2024 Program Topics

CARB TY GIF

Program Topics:

Our Technical Program Committee has developed several program topics for the 2024 ASIC California Conference that they are excited to build. While the TPC is currently deciding initial speaker invitations, we will be opening the opportunity to give a podium or poster presentation to all the researchers, regulators, communicators, industry members and community scientists that are working on improving air quality sensors. 


The potential of low-cost sensors in regulatory applications: how far have we gone

Low-cost sensors (LCS) have enabled characterization of air quality through measurement of air pollution concentrations at unprecedented scales, underscoring their potentials to reshape air pollution regulations for designing and implementing precise and targeted pollution control strategies. Yet, the well-known performance limitations of low-cost sensors have greatly hindered their use in regulatory air monitoring that requires rigorous data quality standards. This session welcomes submissions that explore innovative, groundbreaking or even transformative approaches to incorporating data from less precise low-cost sensors into regulatory frameworks. All aspects of related sensor applications are invited, which include but are not limited to: novel sensor/monitor design, new calibration approach, innovative data fusion/assimilation method, air pollution policy design based at least partially on low-cost sensor data, among others.  Discussion of limitations and gaps in LCS performance that prevent their current use in regulatory monitoring is also encouraged.

Session Chairs:

Raiford Hann, California Air Resources Board, Alastair Lewis, National Centre for Atmospheric Science & University of York, Haofei Yu, University of Central Florida

Presentations:
  • Air Sensors Potential for Regulatory Applications: US National Ambient Air Quality Standards and Other Uses 
  • Presentation by: Karoline Barkjohn, Physical Scientist, US EPA, United States

    In the United States, the Clean Air Act requires the Environmental Protection Agency (EPA) to establish National Ambient Air Quality Standards (NAAQS) and to monitor six common air pollutants. For NAAQS compliance, air quality instruments must adhere to stringent sampling, siting, and quality assurance requirements. The EPA’s Office of Research and Development must evaluate and designate instruments as Federal Reference Methods (FRMs) or Federal Equivalent Methods (FEMs) based on accuracy, precision, freedom from interferences, detection range, drift, and other key parameters. The FEM and FRM review process is resource intensive. Currently, both particulate matter (PM) and gas sensors have limitations that prevent them passing the rigorous FEM/FRM review process. Nonetheless, there is a burgeoning market for air quality sensors that, until recently, have been developed without specific guidelines for acceptable performance levels. In response, the EPA has developed sensor performance targets and testing protocols to evaluate their performance, providing a basic understanding of performance that is much less stringent than the FEM/FRM review process. In addition to variations in data quality requirements, the goals of sensor networks may differ from those of the NAAQS air monitoring network. While sensors are not used for NAAQS compliance monitoring, they can complement regulatory monitoring by offering a better understanding of spatial and temporal variations in local air quality, which is useful for determining the placement of regulatory monitors. The extensive collection of sensor-based air quality observations constitutes a valuable dataset for scientific research, supporting the regulatory process. For example, sensor data can be used to estimate exposure for population health studies, assess and enhance air quality model performance, and identify highly localized air pollution sources (e.g., hot spots or leak detection) that may warrant further rigorous investigation.  Overall, these applications can support the weight of evidence analyses of air pollution health effects that are used in the integrated science assessments. These assessments form the scientific basis of the NAAQS limits. In addition, publicly displayed sensor data provides a means of communicating more personalized air quality information to impacted communities, potentially being more effective in raising awareness and encouraging actions to reduce exposure. 

    Although this abstract was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. 
  • Sensor Adoption in Building Certification Programs: LEED, WELL and Beyond
  • Presentation by: Eric Sun, International WELL Building Institute, United States

    This presentation will describe the adoption of sensor technology in popular building certification programs. It will:
    -    Introduce popular building certification programs (e.g., WELL, LEED, RESET, NABERS, BREEAM). Provide an overview of each program’s mission, covered concepts, scoring system and certification process.
    -    Describe the evolution of sensor technology adoption for each program. Compare how sensors were initially utilized to how sensors are currently used in each program.
    -    Compare and contrast the scope and extent of how sensors are utilized between the certification programs.
    -    Detail the main drivers for sensor adoption (e.g., market demand, epidemics, technology, cost, data).
    -    Describe the current limitations of sensor technology for certification programs and how they can be / are being addressed. Postulate the future of sensor technology in these programs.
  • How a District and Community Groups Can Use 4,000 Air Sensors in the San Francisco Bay Area for Policy Decision-Making to Community Awareness
  • Presentation by: Helena Pliszka, Air Quality Data Scientist, TD Environmental Services, United States

    The Bay Area Air Quality Management District (BAAQMD) has created an extensive outdoor air sensor PM2.5 data set spanning the nine counties of the San Francisco-Oakland-San Jose areas. This data set is being used to support both internal air district decision-making and local community-based organizations. The comprehensive air sensor dataset contains 5 years worth of processed PurpleAir data across over 4,000 sites in the area, with plans to expand to other air sensor devices and companies. This comprehensive dataset has undergone extensive data wrangling and processing led by the Bay Air Center (BAC), a technical community resource supporting local air monitoring and air quality education, provided by the air district. BAAQMD, through the support of BAC, is using this dataset in a variety of applications, such as understanding wildfire pollution spread, investigating the effectiveness of Spare The Air days, describing disproportionate PM exposure across neighborhoods, and examining spatial relationships between sensor and reference monitoring sites. BAC is also distributing the dataset to the public, increasing data availability and supporting community-based organizations in their own air quality projects and decision-making. This presentation will describe these applications in more detail by describing objectives and sharing results. 

  • AirNow Fire & Smoke Map: Building Trust in Air Quality Sensors

  • Presentation by: Ron Evans, Senior Analyst, US Environmental Protection Agency, United States

    Public trust and government air agencies familiarity with sensors are an important factor in applications for air quality sensors.  Over 4 years, the AirNow Fire & Smoke Map has demonstrated that specific sensors could be used concurrently with permanent monitors to provide the public with information to protect their health from wildfire smoke.   While non-regulatory, the AirNow Fire & Smoke map shows how air quality sensors can provide significant public health benefits.  Building public trust and government air agency familiarity with air quality sensors thru applications like the AirNow Fire & Smoke Map are important steps in future applications of these sensors.

    The US Environmental Protection Agency and the US Forest Service partnered to develop the AirNow Fire & Smoke Map and it was released in August 2020.  It now has over 1,000 air quality monitors and 14,000 air sensors.  It provides the United States public with local air quality information to make decisions on how to protect their health from wildfire smoke.  The map contains both air quality monitor and sensor data in real time keyed to the EPA Air Quality Index.  Since release there have been nearly 50 million page views, and Version 4 (to be released prior to conference date) represents a significant enhancement to the information provided to the public.

  • A Decade Using Air Sensors: 15 Insights and 5 Predictions

  • Presentation by: Tim Dye, President, TD Enviro, United States

    The past decade has witnessed remarkable advancements in air sensors in the ever-evolving landscape of air quality monitoring. The air quality community has learned a lot in the past 10+ years about how we can benefit from this new technology and, in some cases, also be challenging. With the increased funding and new people and organizations beginning to monitor air quality, it's important to recap, share insights and collectively look at this technology's future applications and benefits.

    During this engaging and fast-paced talk, attendees will learn many air quality community lessons. The lessons will include insights like the "20/80 rule" – spend less than 20 percent of a project's budget on hardware/software and 80 percent on people, analysis, meetings, and action. These insights would encompass technological breakthroughs, data analysis methodologies, and the socio-environmental impact of air sensors. We will also offer a shared vision for the next five years and aim to predict future trends and challenges. 

    This will be a level-setting and engaging talk for all attendees and help frame where we are and are going as a community of experts, researchers, community members, and academics.

  • Real World Sensor Application

  • Presentation by: Wandji Danube Ngongang, Research Fellow, Kenya

    Air pollution is one of the major risk factors for global mortality and populations in Low- and Middle-Income Countries (LMICs) such as sub-Saharan Africa face particular risks and yet effective management programs have not been fully established, in part due to limited monitoring infrastructure. Even, so the burdens of air pollution are likely to be underestimated because of the limited ground monitoring data in Africa. To tackle the problem of poor air quality and its effect on people’s health, it’s necessary to understand how much air pollution people are exposed to.

    The African context presents present a lot of factors that seem to affect the performance of air quality sensors when set in the real-world environment. Their performance is mainly affected by several factors associated to security of the targeted location, the ambient conditions with real emission sources, poor to inexistent internet infrastructure and unreliable power supply. The above-mentioned factors coupled with environment factors like rain, insects and high dust emissions tend to affect the operation of the sensor leading the data loss. 

    The African context is prone to many challenges that seem to make the operation of some imported sensors not fit for the purpose of sending reliable data unless proper site survey is done prior sensor deployment to understand the environment.
    Designing instruments for the unique conditions of Africa is a critical step towards bridging the data gap in the continent. Other sensors like Sail hero, IQair, Praxis, Kunak to list a few have proven to be effective in delivering stable data in the context of outdoor spaces in Sub-Saharan Africa settings. 

    At the same time, numerous efforts from citizen sensing initiatives, research groups, international non-profit organizations, etc. are underway in Africa to encourage the uptake of low-cost sensors and provide protocol of air quality sensors that is fit and adapted for the context of Africa. 

  • Invited Presentation by: Michael Hannigan, University of Colorado Boulder

  • Presentation title and description forthcoming.

  • Leveraging Low-Cost Sensor Technology in High-Value Air Quality Science and Policy Making Applications

  • Presentation by: Zhi Ning, Hong Kong University of Science and Technology

    In recent years, the development and application of low-cost sensor technology have revolutionized air quality monitoring and management. This presentation will explore the significant advancements in sensor technology that have occurred in the last decade. These innovations have enhanced the precision and reliability of detecting key pollutants, achieving measurements comparable to traditional reference methods with traceability. A critical review of these developments is timely, particularly as we consider their impact on data-driven policy making and intelligent air quality management strategies.
    A key focus of the presentation will be the practical applications of this technology in diverse settings. These include landfill environments, where monitoring is crucial for mitigating environmental impact, and urban areas, where fine-scale air quality modeling is essential for public health forecasting and intervention. The presentation will also delve into the use of mobile networks for managing traffic emissions. This aspect is particularly relevant for urban settings, where traffic contributes significantly to air pollution.
    Overall, this presentation aims to highlight how low-cost sensor technology is not just a tool for measurement but a catalyst for smarter, data-driven approaches to high value air quality management, ultimately contributing to healthier, more sustainable environments.

  • A Case Study Assessing PM2.5 Pollution levels in Somanya and Accra, Ghana, Utilizing Low-Cost Air Quality Monitors.

  • Presentation by: Solomon Lomotey, Lecturer/Research Scientist, UNIVERSITY OF ENVIRONMENT AND SUSTAINABLE DEVELOPMENT, Ghana

    Recently, from the census conducted, the population of Ghana gained rapid growth over the past decades, which poses a significant threat to urbanization in various cities. Particulate matter air pollution poses an imminent risk for various air-borne illnesses that can cause respiratory issues, death, and other related issues in major urban municipalities in Ghana. It is recognized that several meteorological factors, including wind, temperature during rainfall, and relative humidity, influence these pollutants. This study uses Low-Cost air quality monitors to assess the PM2.5 pollution in two major cities, Accra and Somanya. Relative humidity, temperature, and PM2.5 were compared between the two cities using correlation analysis techniques. Our findings show a negative link between temperature and relative humidity in the two cities and a positive correlation between PM2.5 pollution and relative humidity. Furthermore, the study observed diurnal and seasonal variations in the parameters between the two locations, which can be attributable to meteorological conditions. In addition, in 61% of the provided data, the measured PM2.5 values were higher than the WHO 24-hour average Air Quality criteria of 15 µg/m3. The results of this preliminary study are essential for managing and controlling air pollution in regions that were previously unmonitored and had limited capabilities, as shown in broader areas of Ghana and Africa, utilizing relatively Low-Cost Sensors.


Community-Centric Data Collection: Unleashing the Power of Mobile Sensors

In an era marked by growing concerns about air quality and its impact on public health, the integration of mobile air sensors into community applications has emerged as a promising tool. This session aims to explore how mobile air sensors are revolutionizing the way communities engage with and address air quality issues. For this session, we seek to bring together experts, researchers, and community leaders who are at the forefront of leveraging mobile air sensors for community-driven air quality monitoring, to explore how mobile air sensors empower communities with real-time air quality data, foster environmental awareness, and drive informed decision-making.

Session Chairs:

Olivia RyderSonoma Technology, Marwa El Sayed, Embry Riddle Aeronautical University, Melissa LundenAclima

Presentations: 
  • Assessment of Personal Exposure to Particulate Matter on Banana Beer and Brickyards Industry Workers in Musanze, Rwanda using Low-cost Wearable Sensors
  • Presentation by: Abdou Safari Kagabo, Assistant Lecturer and PhD Student, University of Rwanda, Rwanda

    BACKGROUND: 

    Personal exposure (PE) to particulate matter (PM) from anthropogenic activities is a major concern worldwide and in the cities of Rwanda, in particular. 

    METHODS: 

    In this study, PM data from two local industries, banana beer and brickyards Production, were collected continuously from 12th January to 10th June 2023 using low-cost wearable sensors. This study assessed and characterized PE to PM for banana beer and brickyard workers in Musanze. PM statistical values were calculated and time series were plotted for all participants. The PE in terms of PM concentrations and time spent was computed and presented. 

    RESULTS: 

    The diurnal variation of PM concentration shows the highest values in the working hours. The pick for both PM2.5 and PM10 concentrations were observed in the early morning between 03:00 and 09:00 AM for banana beer workers with 113 µg/m3 and 136 µg/m3 for PM2.5 and PM10, respectively, and from 07:00 AM to 16:00 PM for brickyard workers with 153 µg/m3 and 234 µg/m3 for PM2.5 and PM10, respectively. The nonworking hours were characterized by low concentrations. The results reveal that the change in mean concentrations as well as the integrated PE depend highly on the time spent in a microenvironment and the nature of the emission sources present in it.

    CONCLUSION: 

    The exposure seems to be higher at work than in other environments due to the activities generating PM present there. Despite low levels in environments other than work, all mean PM concentrations exceed WHO air quality guidelines for PM, indicating potential health risks associated with prolonged exposures. The study showed the impact of time spent in different MEs and the emission sources present in them. The study findings will contribute to understanding occupational health risks and suggest the need for interventions to reduce exposure to PM in the banana beer and brickyard industries. 

    KEYWORDS: 

    Particulate Matter, Personal Exposure, Banana Beer and Brickyards

  • Evaluating Mobile Low-Cost Sensors for PM2.5 Monitoring: Lessons from Snifferbike Sensor Kits in Urban Settings
  • Presentation by: Amirhossein Hassani, Nilu, Norway

    Municipalities aim to provide clean air in Nordic countries, where around 75% of the population resides in urban regions. However, resource limitations hinder traditional air quality monitoring, particularly in smaller towns. The growing interest in mobile low-cost air-quality monitoring technologies has prompted their use in various research and public engagement initiatives for monitoring air quality. This research aimed to explore how mobile low-cost air quality sensors (LCSs) can support communities in active participation in generating reliable and policy-informing air quality monitoring data. The study was performed within the NordicPATH project, which explores new methods for using data derived from LCSs through Urban Living Labs in the Nordic countries. The NordicPATH project’s overall objective was to establish a new model for citizens’ participation and collaborative planning through applying a co-monitoring system and promoting a more inclusive planning process. The co-monitoring system was to combine environmental measurements from official monitoring stations and citizens’ own measuring devices. Within the Urban Living Lab of Kristiansand (Norway), volunteer participants (N = 10) were provided with Snifferbike sensors (Sensirion SPS30), enabling them to have a personalized assessment of exposure to air quality along their daily cycling routes. They were also invited to participate in co-design and co-creation workshops. Through extensive co-location and intercomparison tests, Machine Learning (ML) techniques, and optimal mobile and static calibration strategies, we improved the quality of outdoor PM2.5 (Particulate Matter < 2.5 μm) concentrations measured by Snifferbikes, making them suitable for generating policy-informing maps. We found calibrating sensors in stationary configurations might not apply to mobile settings, and utilizing data from bikes passing near reference stations is recommended. Our findings also suggest that bike speed contributes to increased uncertainty in low-cost PM2.5 sensor measurements, with an estimated addition of 0.03 - 0.04 μg m−3 to the Standard Deviation for every one km h−1 speed increase. Enhanced measurements were utilized to generate PM2.5 concentration maps along the frequently used cycling routes. We introduced a method to determine the minimum number of required PM2.5 measurements per road segment to address the sampling bias issue. In this case, we recommend a minimum of 27 measurements per road segment. Having more data in a transect increases the accuracy and reliability of mapping. This work contributes to developing a data analysis package for enhancing the quality of a network of citizen-operated mobile LCS measurements and unleashing the power of mobile LCSs for actively engaging communities in the air quality monitoring processes. Active involvement implies direct engagement in monitoring, such as through manual data collection or real-time observations, while the passive approach involves setting up sensors and allowing them to gather data without continuous human intervention. The active approach is more hands-on and requires ongoing participation, while the passive method relies on automated data collection. In addition, mobile LCS monitoring has the potential to inform policy and urban planning with data on citizens’ personal exposure in their daily commute that otherwise would be difficult to access.
  • Advancing mobile air quality monitoring with low-cost sensors: A stop-and-go approach for hyperlocal insights in a coastal city of India
  • Presentation by: Dheeraj Alshetty, Advisor Air Quality Scientist, Environmental Defense Fund, India

    Low-cost sensors (LCS) are increasingly popular for static air pollution measurements, but their potential in mobile monitoring remains underexplored due to challenges associated with data quality and reliability. This study adopted a STOP-and-GO method to address some of the challenges, facilitating separate analysis of data collected during the STOP and GO phases.  Mounted on an electric vehicle, the LCS gathered real-time PM2.5 and NO2 data within a 3 x 3 km study domain, centered around a reference monitor. The 21-day monitoring period on predesignated routes, with three repetitions a day, allowed for comprehensive analyses of morning, afternoon, and evening periods. Segregated and corrected (based on collocation) Stop data were further analyzed, generating high-resolution maps for each pollutant. These maps identified dynamic hotspots (hyperlocal variations) in on-road air pollution levels at a finer scale.
    Hotspots revealed by mobile monitoring pinpointed areas with elevated traffic congestion, construction, and burning sources. The findings underscore the potential of LCS for indicative monitoring, offering valuable insights to address air pollution at a hyperlocal level, particularly in low-and middle-income countries (LMICs) where reference monitors are scarce. This research contributes to the advancement of mobile monitoring, providing crucial tools for managing air quality in dynamic urban environments.
  • Mobile low-cost sensors use for route assessment in Aarhus (DK)
  • Presentation by: Francesco Cappelluti, Consultant, Danish Technological Institute, Denmark

    As a part of the EU-funded DivAirCity project, the particulate matter levels along different routes bordering a trafficked road in the city of Aarhus (Denmark) has been measured to identify the least polluted route and incentivize citizens to use it. For doing so, particle monitors OEMs from Alphasense have been laboratory calibrated against a high-end particle counter, paying particular attention to the exploration of the conditions that the devices were going to experience in this type of application (i.e., abrupt changes in the particle concentration due to presence of crossroads, transit of vehicles, local wind conditions, and so on). Subsequently, the OEMs have been assembled into actual mobile sensors, able to geotag the air quality measurements and send data in real time to the cloud. The sensors have been assigned to citizens recruited among the vulnerable categories, whom are the target of DivAirCity project, such as wheelchair users, who were asked to carry them at respiratory height in their daily movements along the routes focused by the project. The results of the measurement campaign give a deeper understanding of the air quality at microscale and will, among other things, help the Aarhus’ policymakers in improving the quality of life of the city’s vulnerable population groups.
  • Ambient air quality measurements of PM, NO2 and O3 in a near-highway community using mobile monitoring and stationary continuous monitoring
  • Presentation by: Mayra Chavez, Visiting Professor, University of Texas at El Paso, United States

    Mobile air sensors have emerged as a promising tool for community applications, as concerns about air quality and its impact on public health continue to grow. This project evaluated the feasibility of using mobile monitoring on fixed routes for near-road exposure assessment. Mobile measurements of four air pollutants (PM2.5, PM10, NO2, and O3) were continuously recorded using U.S. EPA-certified Federal Equivalent Method (FEM) monitors inside a vehicle traveling repeatedly in a near-road community. Spatio-temporal mobile air quality data were aggregated and compared to data collected at two near-road fixed stations, installed in residences within the community. The first objective assessed the suitability of using the mobile monitoring data to represent community exposures to transportation-related air pollutants (TRAPs). The second objective evaluated the feasibility of using mobile instead of near-road air monitors. Mobile monitors successfully detected PM10 concentrations near-road as well as in the community, which were comparable to the background concentrations. PM2.5 and O3 concentrations detected by mobile monitors were similar to those detected near-road in the community. NO2 concentrations detected by the mobile monitors varied from the near-road monitors due to the interactions with ambient temperature, vehicle emissions, and atmospheric chemical reactions. This study found that community exposures to TRAPs can be represented by short-term spatio-temporal measurements using mobile monitors. Mobile air pollution measurements provide a rapid assessment of the air quality in a community without installing multiple stationary sites.
  • Funding a global community led air quality action network
  • Presentation by: Olivia Sweeney, Senior Specialist - Data, Clean Air Fund, United Kingdom

    Philanthropy is uniquely placed to fund a different approach to tackling air quality. Funders can take more risk, innovate, act dynamically which is why we are well placed to support the development and use of wearable sensors, through community led projects across the globe. 

    Wearable sensors can be used as both information collecting tools, but also as campaigning and storytelling devices. They can be an effective tool to show the co-benefits of air pollution solutions, with other interventions for health and climate, and can be an asset to any organisations may they be focussed on air quality or not. 

    We have supported a small programme of community-driven initiatives. Having funded wearable sensor manufacturers, community organisations and global networks to use wearable sensors, we are beginning to collate best practice and lessons learnt from this type of air quality project. 

    As a global north-based funder we cannot rely on wearable sensors alone to ensure community empowerment. Wearable sensors will not address the inequitable burden of air pollution, they are a tool for doing so, deep equity will come from changes in behaviours and actions, at individual, organisation, and institutional level. 

    This session will show our learnings of grant making to these organisations so far, what these projects have achieved and how they plan to develop. It will explore our future strategy for granting, influencing others to support these projects and share power in doing so.

  • Evaluating the Use of Low-Cost Sensors for Mobile Monitoring
  • Presentation by: Priyanka deSouza, Assistant Professor, University of Colorado Denver, United States

    Low-cost sensors (LCSs) for measuring air pollution are increasingly being deployed in mobile applications, but questions concerning the quality of the measurements remain unanswered. For example, what is the best way to correct LCS data in a mobile setting? Which factors most significantly contribute to differences between mobile LCS data and those of higher-quality instruments? Can data from LCSs be used to identify hotspots and generate generalizable pollutant concentration maps? To help address these questions, we deployed low-cost PM2.5 sensors (Alphasense OPC-N3) and a research-grade instrument (TSI DustTrak) in a mobile laboratory in Boston, MA, USA. We first collocated these instruments with stationary PM2.5 reference monitors (Teledyne T640) at nearby regulatory sites. Next, using the reference measurements, we developed different models to correct the OPC-N3 and DustTrak measurements and then transferred the corrections to the mobile setting. We observed that more complex correction models appeared to perform better than simpler models in the stationary setting; however, when transferred to the mobile setting, corrected OPC-N3 measurements agreed less well with the corrected DustTrak data. In general, corrections developed by using minute-level collocation measurements transferred better to the mobile setting than corrections developed using hourly-averaged data. Mobile laboratory speed, OPC-N3 orientation relative to the direction of travel, date, hour-of-the-day, and road class together explain a small but significant amount of variation between corrected OPC-N3 and DustTrak measurements during the mobile deployment. Persistent hotspots identified by the OPC-N3s agreed with those identified by the DustTrak. Similarly, maps of PM2.5 distribution produced from the mobile corrected OPC-N3 and DustTrak measurements agreed well. These results suggest that identifying hotspots and developing generalizable maps of PM2.5 are appropriate use-cases for mobile LCS data.
  • Real-time Portable Air Pollution Sensor Data for Informing Communities and Guiding Local Development and Funding Decisions
  • Presentation by: Janice Lam Snyder, Program Manager, Sacramento Metropolitan Air Quality Management District, United States
     

    The Sacramento Metropolitan Air Quality Management District (Sac Metro Air District) in partnership with the City of Sacramento deployed 200 portable air pollution sensors in low-income and overburdened communities in the City of Sacramento, California to provide highly localized data and build community awareness and inform local land use planning and clean air project funding decisions. The selection of sensor recipients was based on location to help fill data gaps in areas which included multiple indicators such as communities with the greatest health needs, socioeconomic factors, location of emission sources of toxic air contaminants and greenhouse gases, locations of sensitive receptors, areas lacking community development, income, housing, employment opportunities, transportation, medical services, nutrition, education, and a clean environment. 

    Once communities were identified, sensor deployment was prioritized for schools to support proactive strategies to protect the health of youth during wildfire smoke events and support communities with no available neighborhood-level air quality data. The goal to increase spatial coverage of the air monitoring network to fill information gaps, particularly during wildfire smoke events, was achieved by deploying sensors to residents,  businesses, and city facilities to ensure coverage in key areas. Participants must make sensor data public, enabling other community members to inform personal decisions and local officials to consider the information in general planning and air quality management practices. Additionally, the Sac Metro Air District is developing an educational curriculum for middle or high schools serving the general targeted communities to help promote air quality awareness and using portable air quality sensor data. Lastly, this partnership further augmented existing data through a mobile monitoring campaign with professional-grade monitors to further the understanding of localized air quality conditions.  Information from the partnership will also inform future efforts to deploy air quality solutions, such as tree plantings.

    Ultimately, the project partnership advances community-based advocacy priorities for enhanced air quality monitoring efforts. The unique partnership and local funding also facilitated a significant augmentation of the existing air quality network beyond previous efforts, equipping priority under-resourced communities with localized air quality data. 


Community Ownership and Use of Sensor Data

For air quality sensors to fulfill the promise of helping to improve public and environmental health, data must be effectively communicated appropriately and clearly for a given audience. This means knowing and understanding the audience first and foremost. Effective communication enables data to be used in personal and decision-making, which could inform local actions and policies to improve air quality. To explore this topic, we welcome presentations and posters discussing, but not limited to, the following:

  • Public and community partners who have worked with researchers using sensor data – their experience and learnings from case study projects;
  • Translating technical details (e.g., regarding air quality science or sensor data quality) into information for public audiences and key lessons learnt;
  • Sensor data visualization and best practices.
Session Chairs:

Ashley Collier-OxandaleColorado Department of Public Health, Rowena Fletcher-Wood, Environmental Chemistry Group, Daisha WallCleanAIRE NC

    Presentations: 
    • Democratizing Air Quality Data: Bridging Data Gaps in Low and Middle-Income Countries with Low-Cost Sensors
    • Presentation by: Christi Chester Schroeder, Air Quality Science Manager, IQAir, United States

      Ongoing advancements in low-cost ambient air quality monitors have made them a viable public alternative for countries, regions, and territories that lack government-operated air quality monitoring stations. In 2022, independently operated low-cost air quality monitoring provided the only real-time air quality data for more than 40 low and middle-income countries (LMICs). As nations grapple with the dual challenge of environmental health and economic development, challenges can arise in the interpretation and utilization of air quality data. This presentation will explore some of the unique challenges and opportunities faced in the deployment of low-cost sensors in LMICs and the effective communication of data-driven air quality information.  
    • The Importance of Collaboration and Community Inclusion in Air Monitoring to Resolve Environmental Justice and Policy Issues: A West Atlanta Case Study
    • Presentation by: Gwendylon Smith, Executive Director and Founder, Community Health Aligning Revitalization Resilience & Sustainability, United States 

      Research shows that fence line communities suffer health impacts from exposure to emissions from nearby industry.  The quality data needed to help identify and quantify pollutants and their sources is expensive to obtain, and often must meet extensive monitoring requirements. Regulatory monitoring stations are typically located in urban background locations far from fence line communities and often do not measure all criteria pollutants, black carbon or Volatile Organic Chemicals. This work will discuss the collaborative efforts of two community organizations CHARRS (Community Health Aligning Revitalization Resilience & Sustainability) and WAWA (West Atlanta Watershed Alliance), with private sector, municipalities, State, and higher education institutions (2B Tech, TD Environmental Services, Georgia Environmental Protection Division, Atlanta Fire and Rescue, University of Maryland, and the University of Georgia). Together, these groups are working to address the gap in air quality data for West Atlanta by using 2B Tech’s AQSync monitor and handheld Personal Air Monitors (PAMs). These monitors will collect real-time data to measure local pollution in order for (1) community members to educate themselves on the connection between the built environment, air quality, and health risks; (2) document any disproportionate levels of pollution; and (3) evaluate the efficacy of current emission policies and regulations.
    • Strengthening Media Capacities to Report on Air Pollution: Lessons from Nairobi Earth Journalism Network Training
    • Presentation by: Jackline Lidubwi, Program Coordinator, Internews Network, Kenya

      Air pollution poses a significant threat to human health, the economy, and ecological balance in Nairobi and beyond. Rapid deterioration in air quality is fueled by population growth, industrialization, deforestation, construction activities, and increased vehicular traffic. Recognizing the pivotal role of media in shaping public perceptions and policies around air pollution, Internews, as part of the Clean Air Catalyst project, organized a media training workshop titled "Air Pollution Sources, Impacts and Solutions - How to Report on Them." 

      The workshop aimed to empowered Kenyan journalists with knowledge and reporting skills on air pollution, fostering effective and comprehensive coverage of its sources, impacts, and potential solutions in Nairobi. Through expert presentations, discussions, and a field trip, journalists were sensitized to the science of air quality and the socio-environmental implications of pollution, enhancing their capacity to collaborate with researchers and report on-air pollution in a compelling manner.

      Following the workshop, participating journalists enhanced their knowledge of air pollution sources, impacts, and solutions. Their improved skills in information sourcing have contributed to more effective reporting on air pollution in Nairobi. The subsequent story grant activities have heightened awareness and contributed to public understanding of key pollution sources, catalyzing solutions to reduce air pollution and enhance human health. 
    • Breathe Accra Project: Facilitating data-driven solutions to the air pollution problem in Accra
    • Presentation by: Kofi Amegah, Associate Professor, University of Cape Coast, Ghana

      Utilization of low-cost air quality sensor data can help protect public and environmental health and requires a sound framework for collecting, processing and analyzing the sensor data and effectively communicating the data for local action and decision-making. The Breathe Accra Project understands this reality and the need for data-driven solutions to the air pollution problem in Accra. In this presentation, I will outline the data analytics and communication strategy of the project that is facilitating our community advocacy work and policy engagement activities in Accra for improved air quality for the population. 
    • Community-Led Air Quality Monitoring: Early Learnings and Exciting Opportunities in Colorado
    • Presentation by: Madelyn Percy, Project Coordinator, Education & Community Opportunities Unit, Colorado Department of Public Health & Environment, United States

      Given the years of mistrust caused by unjust environmental policies, many disproportionately impacted communities (DICs) are wary of working with governmental entities to address environmental injustice, while also lacking the technical expertise to deploy appropriate environmental monitoring networks. In Colorado, the Air Pollution Control Division’s Education & Community Opportunities unit is working with a number of communities, including Pueblo, Colorado’s Mothers Out Front, to support the deployment and maintenance of a station - satellite sensor network, as well as working with local schools, libraries, and municipal and county governments to share outreach materials and promote understanding of how to interpret the data. While the network is still very new, our unit, in collaboration with local organizers, has tried to use a number of different outreach approaches (e.g., workshops, community meetings, and collaborating with teachers at local schools) to support data literacy in the community. In this presentation, we will outline the outreach approaches that were used, and share initial information about how different stakeholder groups responded. All of the information that we used, as well as the design for our stakeholder engagement plan, are readily shared. Moving forward, we will replicate our successes in partnership with other communities across the state.
    • Using the data: Love My Air Denver's new program with health clinics 
    • Presentation by: Nancy Fitzgerald, Denver Department of Public Health and Environment, United States

      Love My Air's mission is to provide Denver’s diverse communities with visible, accessible, and actionable air quality information. The community air monitoring program started in 2018 at the City and County of Denver with PM 2.5 sensors in Denver Public Schools and is now expanding to place monitors in local community health clinics. Expanding the program to include local healthcare clinics in its mission opens a new door to engage with community and promote awareness of air quality. In the next two years, Love My Air will be co-creating programming with clinic partners to leverage data from air monitors located at the clinics to improve community awareness of air quality and its impact on health through the influence of trusted community health providers. Each clinic will receive an AQSync monitor measuring O3, NO2, NO, PM 1, PM 2.5, PM 10, CO2, CO, and tVOCs. Real-time data will stream to an interactive kiosk in the clinic lobby along with information about how the various pollutants impact health.  

      Love My Air is discussing the possibility of co-presenting with a community partner that will attend the conference.  Love My Air Denver staff would like to discuss this concept with ASIC organizers.

    • Living and Learning with Air Pollution
    • Presentation by: Savannah D'Evelyn, Postdoctoral Scholar, University of Washington, United States

      The burden of air pollution on children’s health is increasing as exposure to wildfire smoke and other airborne pollutants is becoming an inevitable yearly occurrence. Children are one of the most at-risk populations for respiratory health effects from exposure due to the fact they often spend more time outdoors, they breathe in more air per minute, and their lungs — the organ most readily exposed to airborne pollutants — are still developing. School environments provide a location for smoke exposure interventions that, during the school day, could give all students equal access to clean air, as well as an opportunity to engage students as advocates for smoke exposure prevention within their families and the larger community. The overall goals of this project are to reduce exposure to air pollutants in school settings thus improving student health, and to promote health equity by addressing barriers to clean air. Through Youth Participatory Action Research, four high schools across Colorado have set up student-run monitoring networks to evaluate air quality within their school communities. Students are currently in the process of analyzing monitoring data and will then move into the co-development of intervention plans to improve air quality and reduce exposure to air pollution. 
    • Enhancing Air Quality Communication in African Cities: Experiences from C40’s AC4CA Communications Training Workshop
    • Presentation by: Zoe Chafe, C40 Cities

      Effective communication of air quality information is crucial for supporting ambitious action, raising awareness, and achieving air quality goals. C40’s African Cities for Clean Air (AC4CA) initiative conducted a three-day communications training workshop to equip air quality technical and communication staff from African cities with the skills and knowledge to translate technical information into accessible messages for stakeholders and the public.

      The workshop built upon a series of four preparatory webinars covering air quality communication needs assessment, Air Quality through Urban Actions (AQUA) tool training, translating AQUA data into communication, and media landscape analysis. The in-person workshop delved deeper into these topics and provided practical hands-on exercises to enhance participants' communication skills.

      The workshop proved to be a valuable resource for participating cities, empowering them to effectively communicate air quality information and advance their clean air goals. The workshop's emphasis on practical skills and tailored communication plans will undoubtedly contribute to improved air quality outcomes in African cities.

    • Targeted Data Visualization
    • Presentation by: Jonathan Callahan, Associate Research Professor, Desert Research Institute, United States

      There is no one-size-fits all for data visualization of air quality data. Yes, maps and time series are important. But every community will need to accommodate members who are color vision impaired or who are non-English speakers. Every community will have members who are interested in the details and others who just want a summary assessment. Every community will have a range of decision making needs: When should I walk the dog? Should we cancel school sports? Is this sensor functioning well enough to believe its data? Those communicating air quality data with local communities need to have a range of options when processing and displaying that data. This talk will present a suite of data visualization options that have been used in real-world community outreach projects.
    • Air Justice Lab: Community Air Monitoring in the Capital Region of New York 
    • Presentation by: A'Livija Mullins-Richard, Project Lead for Air Justice Lab, The Sanctuary for Independent Media, United States

      The Air Justice Lab (AJL) is a community air monitoring and educational project based in Troy, NY at NATURE Lab, a community health and urban ecology initiative operating out of the Sanctuary for Independent Media. AJL aims to identify, increase awareness and action on air quality issues in the Capital Region. The project utilizes low-cost PurpleAir sensors installed throughout the area with a focus on characterizing fine particulate matter (PM2.5) outdoors from factors such as high traffic volume and proximity to toxic waste sites. With 50 PurpleAir sensors, AJL aims to supplement and collaborate with other environmental justice networks to further quantitative credibility in the efforts to guide action and understanding air quality issues. Nearly half of the monitors installed by AJL are located at homes that are neighbors to major emitters like Norlite and LaFarge. Sensor data are corrected using local calibration measurements by co-locating sensors with a reference monitor. The quality-controlled data are being used to determine spatiotemporal variability and to identify local source impacts. The data collected will be presented with an interdisciplinary approach that bridges science and art to effectively engage community members in addition to revealing decades of environmental inequity in the Capital Region.

    Challenges and successes in community air quality monitoring: implementing and evaluating community-based air quality monitoring programs

    Community scale monitoring programs have been rolled out in dozens of locations over recent years because of advances in air sensor technology. New sensor technologies have the potential provide insight into air pollutant concentrations and disparities on neighborhood scales, ultimately yielding valuable information to improve public health. Understanding past challenges and successes is key to building the most effective community monitoring programs in the future. In this session we will explore existing community air monitoring systems with a focus on (1) how new technologies and methods are being applied to tackle monitoring objectives, (2) how programs have complemented and expanded on regulatory monitoring (3) how data has been leveraged to create action, and (4) how program success has been evaluated quantitatively.

    Session Chairs:

    David RidleyCARB, Amanda KaufmanUS EPA, William Porter, UC Riverside

    Presentations: 
    • Household-Heating Induced PM2.5 Levels Association with Meteorology: Insights from a Community Operated Low-Cost Sensor Network
    • Presentation by: Amirhossein Hassani, Nilu, Norway

      Poland faces severe air pollution challenges, mainly from using coal and fossil fuels for household heating. These pollutants have significant adverse impacts on public health and the environment. Reliable air quality data are crucial to guide policies promoting cleaner energy alternatives. However, Poland’s air quality monitoring infrastructure out of dense urban settlements is limited, with few reference stations and outdated equipment. Employing a network of low-cost sensors managed by local residents, this study focuses on the spatio-temporal dynamics of Particulate Matter (PM) generated by household heating. The research objectives additionally included developing a data quality assurance framework for PM sensors, estimating Relative Humidity-induced uncertainties in sensor data without co-location, and constructing an interpretable Machine Learning model — Generalized Additive Model (GAM) to investigate the relation between meteorological factors, such as temperature, humidity, and wind speed, on PM2.5 levels, especially during periods of high household-heating emissions. Our study was primarily centered on Legionowo, a town selected as the pilot location within the "GREEN HEAT – Towards Collaborative Local Decarbonization" project (https://greenheat.kezo.pl/en/, accessed in October 2023). In this specific town in Poland, our sensor data were predominantly collected through a chosen Pilot Case implemented within a local community, focusing on Legionowo unique characteristics and needs. Data from a network of 13 citizen-operated low-cost air quality sensors (Airly — Airly-GSM, SP. Z o.o., Poland), one official air quality station, and meteorological data were used. Interested local citizens strategically positioned the sensors near places where solid fuels are burned, like in people’s backyards. Key findings include the stable performance of low-cost sensors over the analysis period, with 95.46% of sensor-hour measurements meeting data quality standards. Without any co-location, for PM2.5, we found that there was about a 30% chance of observing a negative bias, meaning the sensor readings were lower than those of the reference instruments across all RH conditions. In the RH range of 50-60%, there was a 66% chance of a positive bias; in the highest RH range of 90-100%, there was a 78% likelihood of a positive bias. Controlling other environmental factors, the Local interpretable model-agnostic explanations (LIME) analysis of the trained GAM revealed that for every 1-degree Celsius rise in air temperature and every one km hour-1 increase in wind speed, the PM2.5 levels in Legionowo would decrease by 0.26 μg m−3 and 0.14 μg m−3, respectively. Conversely, a 1% increase in Relative Humidity (RH) is associated with an increase of 0.03 μg m−3 in PM2.5 levels, respectively. The study’s methodology is transferable and can inform emission reduction strategies and the transition to cleaner heating sources in various Polish urban centers. It underscores the importance of understanding complex interactions between environmental factors and PM pollution, going beyond predictive modeling.
    • Community-based exploration of environmental and human health impacts of industrial air pollution in Pascagoula, MS 
    • Presentation by: Caroline Frischmon, Graduate Student, University of Colorado Boulder, United States

      After nearly a decade of environmental justice organizing in their community, residents of a fence line neighborhood in Pascagoula, MS are still met with skepticism from local officials over whether local industrial pollution impacts the residents’ health. The neighborhood organizing group, Cherokee Concerned Citizens (CCC), sought out a partnership with university researchers to gather data that would quantitatively back their testimonies and convince local officials of the need for environmentally protective action.

      Using community-based participatory research (CBPR) principles, our university-community team deployed multi-pollutant, lower-cost sensors in the neighborhood. The pollutants measured (H2S, SO2, tVOC, NH3, and PM10) are based on the specific odor and dust complaints reported by the community. As the sensors collect data, we asked residents to track their pollution-related health symptoms and perceived industrial odors. In community analysis workshops led by CCC, participants are invited to analyze their own symptom and odor records alongside the concurrent air quality data. Workshop participants then develop stories informed by the overlaps in pollution episodes, odors, and health impacts. These narratives will be shared with local officials.  

      Through this project, we highlight how qualitative community data, such as symptom and odor reports, can narrate air quality data, making meaning beyond typical regulatory and scientific frameworks of low-cost sensing. By employing CBPR in each step of our study, we demonstrate an accessible strategy for communities to gather, analyze, and share actionable air quality data that corroborates their lived experience and supports their environmental advocacy.
    • Progress in community air quality projects: from sensors to action 
    • Presentation by: Ethan McMahon, President, EM Environmental Solutions, United States

      Many air quality sensor efforts have focused on selecting and using air quality sensors and understanding the air quality in cities. Some have engaged with communities to understand their concerns about air quality. This foundational work is important, but very few projects have progressed to the point of action and improved human health, which is why people start these projects in the first place. This presentation looks at the effort that has been dedicated to various phases of air quality projects and will identify potential ways to overcome barriers to action such as expertise, political will and funding.
    • A GAO Technology Assessment of Air Quality Sensors: Promising Technology Faces Challenges to Implementation 
    • Presentation by: Evonne Tang, Senior Biological Scientist, US Government Accountability Office, United States

      Lower-cost air quality sensors have the potential to help close gaps in current air quality monitoring, but challenges with implementing sensor networks and managing data could hinder their usability. GAO conducted this technology assessment in light of congressional interest in the potential role of lower-cost air quality sensors in monitoring air quality. Our recent report discusses (1) what sensor technologies are; (2) what their benefits and uses are; (3) how well they perform and factors that affect their performance; (4) challenges to using sensors; and (5) options policymakers could consider to help address these challenges. 
      We conducted a literature search, held a two-day meeting of subject matter experts, and interviewed stakeholders across sectors to identify challenges with developing and using sensor technologies. For example, sensors for air toxics could be valuable for filling current information gaps, but relatively few sensors are currently available to measure air toxics.  We also identified options policymakers may consider if they wish to address these challenges and advance the development and use of air sensor technology. This presentation will focus on our identification of policy options and set the stage for a panel and audience discussion of their potential implementation.
    • Challenges and successes over 9 years of New Zealand’s Community Air initiative 
    • Presentation by: Ian Longley, Principal Scientist, NIWA, New Zealand

      Since 2015, over 20 community-based air monitoring projects, involving hundreds of community volunteers, have been conducted across New Zealand by a small research team, all based on exploring and expanding the potential of low-cost air sensors. Studies have mostly focussed on winter home heating emissions, and included outdoor and indoor monitoring, school engagement and student participation, survey tools and intervention studies. 
      This presentation reviews the highlights and lowlights of our accumulated experience, some of our best ideas, some of the things we’d scale up if we had the money, and some of the things we will never do again.
    • Lessons learned by a small sensor manufacturer from a decade of community monitoring projects around the world 
    • Presentation by: John Downie, Technical Business Development Manager, Environmental Instrument, United Kingdom

      This presentation provides a manufacturer’s view of a wide range of community projects, from the ground-breaking European Citi-Sense project a decade ago, to City of Minneapolis today. Challenges – and solutions – discussed include getting agreement from stakeholders to mount equipment, sourcing and maintaining a power supply, and the impact on sensitive equipment of increasingly EMF-dense monitoring landscapes. Network calibration solutions are discussed, data management and public engagement. Mixing of air, local vs regional pollution sources and inevitably finite budgets create challenges regarding positioning of monitoring nodes, so whether chosen measurement points are relevant and fair to the communities involved. Projects have informed change at political, infrastructure and environmental justice levels, with data managed and disseminated in many ways. 

      We talk about innovations in technology and its application, but also the tension between the temptation to adopt cutting edge technology and the pitfalls of putting inadequately developed monitoring systems into the hands of community monitoring teams. From supplementing regulatory monitoring in well-instrumented neighbourhoods, to simple projects without regulatory monitoring in developing countries, projects consistently throw up surprising results and outcomes to be understood and learnt from. The challenges of maintaining a programme over time are also discussed: ongoing support, re-calibration, replacing consumables, re-engaging with stakeholders.

    • Community Participation and Scientific Approaches in Air Pollution Monitor Network Siting: The Case of the Community Air Sensor Network of Kintampo Health Research Centre (KHRC), Ghana 
    • Presentation by: Mohammed Mujtaba, Senior Research Fellow, Kintampo Health Research Centre, Ghana

      Global public health is at risk from air pollution, and interest in air monitoring has grown as low-cost air sensors become widely available. Although community involvement has always been essential to health initiatives, few studies in low and middle-income countries show community participation in air quality monitoring. KHRC and its U.S. collaborators used a collaborative, community-engaged process to develop community air monitoring networks that attain the scientific rigour required for research while also providing schools, government agencies, and others access to air quality data so as to be able to address air pollution and climate change and its impacts. This project aimed to outline strategies for involving communities in air quality issues and to list the results that have been achieved through the establishment of rural and peri-urban communities air monitoring networks that together include 60 PurpleAir, 7 AirNote 35 Modulair-PM (PM2.5 and PM10) and 5 Modulair (6 gases, PM2.5, PM10) by Quant-AQ, and a federal equivalency monitor for PM2.5 (Met-One BAM 1022) for particulate matter (PM) monitors in the middle belt of Ghana. 

      This presentation will provide the benefits and challenges of the community-engaged cooperative best practices we have developed. Since Sept 2016, KHRC has involved communities in understanding goals for and siting low-cost sensor networks via the following activities: (1) establishing equitable partnerships with KHRC; (2) forming community steering committees to guide project activities; (3) engaging residents to determine monitor locations; and (4) organizing community meetings to guide display and dissemination of research findings; 5) incorporating community input. Due to the active community involvement in the study, participants in the community have greater awareness, and knowledge, and help protect the infrastructure for research. These strategies resulted in a community air monitoring network that can assist with research activities, direct governmental initiatives, and enhance public health.

    • Challenges and Real-World Lessons from Multiple Sensor Network Implementations in the South Coast Air Basin 
    • Presentation by: Randy Lam, Air Quality Specialist, South Coast AQMD, United States

      Addressing the obstacles that arise from managing sensor networks is vital to fostering trust in a community monitoring project and maximizing the ultimate impact of a sensor network as a whole. This talk will share best practices and lessons learned from the implementation of multi-year sensor networks ranging from 5-100 devices in collaboration with government, community organizations, industry, and academic groups for sensor networks within the South Coast Air Basin. This talk will also discuss the challenges faced with maintenance, community engagement, data dashboards, calibration, and QA/QC in general during the implementation of AB617, SEJCA/EJG2G, and other sensor network projects. 
    • Brightline’s Air Quality Monitoring Network: How Open Data Empowers Communities 
    • Presentation by: Trinity Vang, Policy and Community Organizer, Brightline Defense, United States 

      San Francisco residents and community groups are concerned with the air quality, especially around highways, congested intersections, and industrial sites.  Residents of San Francisco’s Disadvantaged Communities and Priority Populations have shared anecdotes of their windows and park handrails being covered in a film of black soot. The city has only one air quality reference site, which means that people have limited access to localized air quality information to support their lived experiences of bad air quality. 

      To address this air quality data information gap, in 2020 Brightline established a community-driven air quality monitoring program with a number of community partners in eastern and southeastern San Francisco. Together, we deployed 19 low-cost air quality sensors in low-income areas of San Francisco to provide localized, real-time free air quality data. More recently, Brightline has expanded our monitoring to include not only PM2.5 but black carbon and NO2. 

      Using our network, Brightline has put on various air quality community workshops, trained community leaders in air quality data collection, and empowered community members to advocate for cleaner air in underserved San Francisco neighborhoods. Brightline has bridged relationships between the Bay Area Air Quality Management District, Single Room Occupancy Tenant Leaders, and youth leaders across San Francisco. 

      Over the last three years, Brightline’s network has allowed our organization to collaborate with our partners to:
      - Promote air purifier advocacy and distribute device in DACs
      - Expand our network to monitor indoor air quality in Single Room Occupancy units
      - Expand our network to SouthEastern San Francisco communities such as Bayview Hunters Point
      - Inform San Francisco’s monolingual communities about air quality resources

    • The Richmond Air Monitoring Network: Spatiotemporal Trends, Challenges, and Lessons Learned
    • Presentation by: Boris Lukanov, PSE Richmond

      The Richmond Air Monitoring Network (RAMN) stands out as the first high-density community air monitoring network to collect continuous measurements of four important air pollutants—PM2.5, NO2, ground-level ozone, and black carbon (BC). Measurements were taken at high spatial and temporal resolution across 50 locations in Richmond, California, over a two-year period. The four air pollutants were chosen based on their known impacts to human health and the current state of low-cost air sensor technology, while sites were selected through an extensive community outreach process to complement a sparse network of regulatory monitors in an area home to a variety of air pollution sources and some of the most environmentally and socioeconomically burdened communities in California.

      RAMN was able to expand the region’s access to hyperlocal air quality data and provided granular insights into the sources and patterns of local air pollution. The combination of four key air pollutants and their complex spatiotemporal dynamics revealed that BC may be a better indicator for local pollution sources than PM2.5. Besides findings, which help illustrate how a dense sensor network can address current limitations in the spatial coverage of government air monitoring, we will also discuss data quality assurance protocols used, challenges encountered, and lessons learned that, together, help provide a framework for future low-cost air monitoring networks.
    • Impact-driven, stakeholder-partnered air justice research in Boston, MA: 7 years of insights from Air Partners
    • Presentation by: Scott Hersey, Associate Professor and Air Partners Lead, Air Partners & Olin College, United States

      Since its founding in 2017, the Air Partners group at Olin College of Engineering has supported over a dozen partners to conduct projects related to advancing air justice in Environmental Justice (EJ) communities in the Boston area. These organizations have spanned a wide variety of identities, including city governments, community-based nonprofits, EJ advocacy organizations, and elected officials at the city, state, and federal levels. Throughout these projects, several overarching principles have guided our work: 1. Continually decenter our (faculty/student/academic system) priorities in order to center community priorities; 2. Meaningfully involve partners in decision making about study design, implementation, outputs and endpoints; 3. Bias toward impact as the primary goal guiding the creation of outputs from projects. In this presentation we will introduce a roadmap that summarizes our approach to stakeholder-partnered projects, highlighting guiding questions and goals for each stage of a project. We will also share case study examples from partnered projects to highlight lessons learned over 7 years of partnering in air quality research to achieve impact.
    • Community Air Monitoring at the Otay Mesa Commercial Truck Port of Entry
    • Presentation by: Edmund Seto, Professor, University of Washington, United States

      The Caltrans-funded Otay Mesa study expanded a community air monitoring network along the US-Mexico Border to sites that were selected to assess the changing impacts of truck traffic at the commercial port of entry.  This study was accomplished through a partnership between San Diego State University (SDSU), University of Washington, and Casa Familiar, San Ysidro, CA. The network of sensor nodes collected measurements of particle (PM2.5 and black carbon) and gaseous pollutant (NO, NO2, CO) measures at 10 new sites for more than a year. Some sites were installed with the assistance of Caltrans on their property (right-of-way).  As part of the deployment, all nodes were collocated, and their measurements calibrated, to reference instruments operated by the San Diego Air Pollution Control District (SDAPCD). Calibrated pollutant measures were compared to reference instruments operated by SDAPCD at the Donovan site, as well as to reference instruments operated by SDSU at two sites (Tijuana Estuary and San Ysidro, CA). PM2.5 and BC concentrations were higher in all seasons at the Otay Mesa study sites compared to the background SDSU site located at the Tijuana Estuary. One site in particular (the CHP truck facility) had notably higher BC concentrations than the others. Additionally, across most sites and seasons, there was a clear diurnal pattern to particle pollution with the highest concentrations observed during morning hours. Pollutant measures were also found to be correlated with truck-specific traffic flows obtained from the California Performance Monitoring System (PEMS). Findings were presented prior to the final project report to incorporate the feedback of community members. This phase of the study established the important relationship between truck traffic at the Otay Mesa border crossing and air quality, and the need for ongoing monitoring at this highly traffic impacted region.

    Effectively converting air quality data to actionable air quality information: Data science tools to scale QA/QC

    One of the most challenging aspects of running an air quality sensor network is consistently converting real-time, highly spatially-resolved data from a distributed network to information.  Transparent data handling processes that scale in support of local air quality monitoring infrastructure are critical to improving the adoption and use of  air quality information.  In this session we welcome abstracts that explore the development of (1) data science tools around QA/QC for sensor networks, (2) data visualization approaches that enhance dissemination of air quality information, (3) continuous QA evaluation techniques exploring changes in sensor performance over time, including bias and (4) data management, calibration, and control strategies. 

    Session Chairs:

    Alena BartonovaNILU Climate and Environment Research Institute, Eben CrossQuantAQ, IQ MeadImperial College

    Presentations: 
    • Air Sensor Quality Assurance Workshop Summary
    • Presentation by: Karoline Barkjohn, Physical Scientist, US Environmental Protection Agency, United States

      In July 2023, the US Environmental Protection Agency hosted a workshop bringing together experts to discuss air sensor quality assurance (QA). The workshop focused on particulate matter (PM) sensors, fenceline and near-source community volatile organic compound (VOC) sensors, and gas sensors measuring ozone, nitrogen dioxide, and carbon monoxide. These three types of sensors are in different places developmentally with PM sensors being used to provide higher spatially and temporally resolved data with a common set of QA procedures, VOC sensors in use for regulatory purposes in some places (e.g., Colorado), and gas sensors proving helpful for some applications but still needing more work to consistently provide accurate data. Sensor applications discussed included near-source, ambient air monitoring and management, community benefit and environmental justice, validation of satellite and models, indoor use, and research purposes including health and climate research. Different levels of QA will be required for different applications and desired outcomes. Many common QA practices were discussed including field collocation, harmonization, calibration, correction, data analysis methods, and siting. Sensor limitations were also discussed including sensor lifetime and degradation, bias, precision, and the influence of interferent pollutants and meteorology (e.g., temperature, relative humidity). This talk will provide a summary of the 2023 workshop and resources created from the findings.
      Although this abstract was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. 
    • Smart multi-sensor calibration of low-cost particulate matter monitors
    • Presentation by: George Castelar, Head of the Air Quality Department, Metropolitan Municipality of Lima, Peru

      A variety of low-cost sensors have recently appeared to measure air quality in Peru, making it feasible to face the challenge of monitoring the air of large urban conglomerates at high spatial resolution. However, these sensors require a careful calibration process to ensure the quality of the data they provide, which frequently involves expensive and time-consuming field data collection campaigns with high-end instruments. My current investigation proposed machine-learning-based approaches to generate calibration models for new Particulate Matter (PM) sensors, leveraging available field data and models from existing sensors to facilitate rapid incorporation of the candidate sensor into the network and ensure the quality of its data. In a series of experiments with two sets of well-known PM sensor manufacturers, I found that one of our approaches can produce calibration models for new candidate PM sensors with as few as four days of field data, but with a performance close to the best calibration model adjusted with field data from periods ten times longer.
    • QUANT: an overview of a long-term study for the evaluation of commercial air quality sensor systems
    • Presentation by: Sebastian Diez, Researcher, Universidad del Desarrollo, Chile

      Innovations in lower-cost sensor technology provide an opportunity to enhance our understanding and ability to manage air quality issues. While the benefits of increased spatial coverage and real-time measurements are evident, challenges persist regarding sensor reliability and data quality. These challenges become more intricate in commercial implementations, where intellectual property constraints often lead to "black boxes", limiting the user's understanding of the data production process. Here we present an overview of the QUANT (Quantification of Utility of Atmospheric Network Technologies) study, a 3-year assessment across various urban environments in UK. QUANT stands out as one of the most comprehensive studies on air quality sensors conducted to date, remarkable for encompassing a wide variety of companies in a single evaluation, including two generations of sensor technologies. Integrated into an extensive data set, the study provides a long-term appraisal of the precision, accuracy and stability of a range of sensor systems. Findings reveal significant variance between devices, substantial impacts of calibration, challenges posed by sensor relocation in different settings, and the critical issue of long-term stability. The data and outcomes from the QUANT study, open to all, serve as a valuable resource for those looking to implement air pollution sensors, offering insights into their real-world capabilities.
    • The Community Air Quality Viewer (AQview)
    • Presentation by: Raiford Hann, Air Resources Engineer, United States
       

      The California Air Resources Board (CARB) recently developed the Community Air Quality Viewer (AQview), an innovative cloud-based data management system, to be the centralized repository for storing, visualizing, interpreting, and facilitating access to community air monitoring data. AQview accommodates data from monitoring initiatives under California’s Assembly Bill 617 Community Air Monitoring Plans, CARB's Community Air Grants Program, and independent studies. Given the recent growth in the use of air quality sensors, AQview is designed to handle data of varying quality from a wide array of providers, monitoring technologies, monitored pollutants, and quality assurance approaches.

      To address the diversity in the data ingested, AQview conducts robust quality control (QC) checks to flag questionable or invalid data records. AQview ensures the transparency of data processing by making the results of QC checks available alongside each data point in the downloaded files. Data after QC checks are accessible through AQview's intuitive and mobile-friendly visualization tools, including a real-time PM2.5 air quality map and a time-series tool for analyzing recent trends at multiple sites. 

      This presentation offers a comprehensive overview of AQview’s existing features, including its automated and configurable QC checks and data visualization tools. It highlights how these elements furnish community members and other users with actionable air quality data and information, demonstrating the power of transparency and accessibility in air quality management.

    • Enhancing Air Quality Monitoring with a Multi-Step Data Quality Framework for Low-Cost Sensors

    • Presentation by: Amirhossein Hassani, Senior Scientist, NILU, Norway

      As low-cost air quality sensor networks continue to expand, there is a growing need to advance how we utilize the collected data. The current paradigm often involves isolated data analysis at sensor locations, limiting our ability to understand air quality comprehensively. In response to this challenge, this study introduces a novel approach that jointly exploits data from multiple low-cost sensors within the sensor.community database with satellite remote sensing and output of air quality models to improve the quality of the sensor network data. Our research focuses on leveraging a multi-step framework for flagging and correcting sensor data, aimed at enhancing the reliability of low-cost sensor network observations while reducing the requirement for observations from reference-grade instruments. In the first step of our framework, we perform a thorough consistency check of individual sensor attributes, including sensor coordinates. This initial quality control step ensures that the sensor data aligns with the expected geographical locations and provides a foundation for reliable analysis. The second step in our approach is dedicated to identifying periods of sensor measurements with zero variance, which often indicate constant value measurements. Detecting and flagging such periods is crucial to prevent the inclusion of erroneous data in subsequent analyses. The third step of our framework assesses the spatial cross-correlation of each sensor with its neighboring sensors. This step helps identify anomalies or outliers in the sensor network, further enhancing the overall data quality. We utilize Terra and Aqua MAIAC Land Aerosol Optical Depth (AOD) daily data at a spatial resolution of 1 km (Lyapustin and Wang, 2022) and the Copernicus Atmosphere Monitoring Service (CAMS) European air quality reanalysis (approx. 10 km resolution) to quantify the regional spatial correlation. Temporal correlation evaluation forms the fourth step of our approach. Here, we analyze the temporal relationships between sensor measurements, ensuring that the data follows expected patterns over time. Deviations from these patterns are flagged for further investigation. In the final step, we focus on sensors that are susceptible to sensor drift. This step assumes the availability of adequate data to establish a reliable baseline. Sensors showing significant deviations from the established baseline are identified and flagged. The output of our multi-step framework provides a refined dataset with improved data quality. We intend to leverage this continental-scale, high-quality sensor dataset to evaluate its added value in the conversion of satellite-based Aerosol Optical Thickness (AOT) observations from the Visible Infrared Imaging Radiometer Suite (VIIRS) instrument to surface PM2.5 concentrations using the XGBoost machine learning technique and a variety of spatial predictor variables. By incorporating reliable, low-cost sensor observations in the model’s training, we aim to increase the accuracy and reliability of continental-scale surface PM2.5 mapping and general air quality monitoring, which is critical for various environmental and public health applications. This research contributes to the advancement of air quality assessment by integrating the potential of low-cost sensors with established remote sensing techniques. 

      Lyapustin, A., Y. Wang. MODIS/Terra+Aqua Land Aerosol Optical Depth Daily L2G Global 1km SIN Grid V061. 2022, distributed by NASA EOSDIS Land Processes Distributed Active Archive Center, https://doi.org/10.5067/MODIS/MCD19A2.061. Accessed 2023-10-12.

    • Using sensor networks for source apportionment of PM2.5 and investigating spatial representativeness of AQ network stations

    • Presentation by: Stig Hellebust, Lecturer, University College Cork, Ireland

       High density air quality sensor networks offer a unique opportunity to better understand the hyper-local and temporal patterns of PM2.5 in urban settings. This work explores four separate low-cost air quality sensor (LCS) networks in Ireland; a network of devices in Cork city (Ireland’s second city), A network of devices in the Port of Dublin (the Capital), a network of devices in a small inland town impacted by domestic solid fuel burning during the Winter season and a network of LCS in a coastal small town on the South Coast. The LCS networks investigated incorporate devices from Purpleair, Clarity, Quant-AQ and Airscape.
      The data from all networks was subject to time-series analysis with a view to investigate the local variations and temporal trends in pollutant concentrations in all locations. The work explores methods and tools to extract and visualise information on local source contributions by leveraging the high-frequency nature of the data. The high spatial resolution nature of LCS network data was also leveraged to investigate long-term persistent differences in local air quality both within a town and between towns. This helped improve spatial classification of location types, which ultimately will aid monitoring-based assessment of population exposure.

    • Understanding site-specific black carbon emissions using kernel smoothing sectoral analysis

    • Presentation by: Collins Gameli Hodoli, Postdoctoral Research Associate, University of Georgia, United States

      Black carbon (BC) is a component of particulate matter with wide ranging impacts including association with adverse health outcomes, reductions in agricultural productivity and climate change. This study investigated the suitability of kernel smoothing sectoral analysis to extract source features of BC pollution using wind data and BC measurements from 2 instruments (MA350, Aethlabs) alongside Macon Highway, Athens from October 18-30, 2023. With machine learning, we applied Pearson’s correlation analysis for reproducibility and kernel smoothers to weight BC concentrations to its proximity to defined intervals of wind components for source identification. Mean concentration was 1 µgm-3, with highest BC pollution observed on Saturdays (1.75 µgm-3) potentially tied to increased vehicular activity. We observed a high consistency between instruments (r=0.97). Fossil fuel-related (BCff) and biomass-related BC (BCbm) concentrations were 0.50-0.85 and 0.10-0.45 µgm-3, respectively and were influenced by wind directions from the highway. A clear diurnal pattern of BC pollution was observed with peak hours in the late morning and evening, but lower between 12:00 and 18:00 hrs (below 0.5 µgm-3). Our preliminary results show the applicability of a new approach to BC source apportionment aimed at short-lived climate pollutants. Additional analyses will incorporate further assessment of emissions from other local sources.

    • Separating local and background pollution signals from observations in real time using digital filtering.

    • Presentation by: Arón Jazcilevich, Researcher, Universidad Nacional Autónoma de México, México

      In many air-quality studies it is important to separate far away from near pollution sources. Using a digital passband Finite Impulse Response (FIR) filter, an instrument is constructed such that an observed pollution signal is separated into its local and background components in real time. The FIR passband cutoff frequency is selected to isolate the background signal. The local signal is then obtained by subtracting the background signal from the observations.

      The instrument to be demonstrated is built feeding the 1 Hz sample acquisition frequency signal from a low-cost PM2.5 sensor into a laptop running a FIR filter with 1024 coefficients. A graphic window shows the signals as they separate. Among other applications, this separation technique allows the apportioning of local road emissions in real time. Other primary emission detectors such as for CO can be used for the same purpose.

      The operation of the instrument will be explained. Its implementation on the field and application examples will be presented.

    • R shiny Tool to process locally harvested Air Sensor SD Card Data.

    • Presentation by: Sebastian Meledina, Environmental Specialist I, New Jersey Department of Environmental Protection Bureau of Air Monitoring, United States

      Quality Assurance/Control is becoming increasingly important in community-based air monitoring projects. However, access to low cost and easy to use tools for performing QC tasks is limited. There has been a growing interest in using open-source programming languages to develop simple applications for data processing and correction allowing the ease of use and sharing across projects. The Purple Air Data Merger utilizes the statistical programming language R and Shiny to create an interactive GUI (graphical user interface). This application can be used to process locally harvested data into a tidy format while also correcting the data using an EPA derived correction equation. The application is shared through an open-source license on the project sharing platform GitHub. GitHub allows community members to have easy access to the application along with updates and project contributions. The Purple Air Data Merger application has been used to assist in community-based projects in a variety of local air projects in the state of New Jersey.  This project highlights the assistance that community members with limited technical skills may need to perform QC related tasks. Moving forward the creation of this tool can be used as a blueprint to help achieve data quality objectives for community group projects. 

    • A Community-Engaged Process Toward Cost-Effective Solution-Centered AQ Sensor Network Design and Operations
    • Presentation by: Amy Mueller, Associate Professor, Northeastern University, United States

      The health implications, and associated disparities, of air pollution in urban environments are well recognized, yet analyses conducted using low spatial resolution regulatory monitors may not provide sufficient granularity to recommend appropriate interventions (or to evaluate effectiveness once they are implemented). Cost effective sensor-based air quality monitoring solutions have the potential to bridge this gap but still suffer from data quality issues such as seasonal bias, weather-based anomalies, and/or poor precision due to temperature and relative humidity effects. Overcoming these data quality issues is critical to (1) ensure a high level of accuracy in data products and conclusions drawn from sensor data, (2) gain stakeholder trust, and (3) develop sensor network design tools that will enable municipalities to achieve their air quality goals in a cost effective way. However, rather than looking to more costly hardware solutions, our team is working together directly with stakeholders to assess the level of accuracy and precision actually needed to answer core municipal and community questions, therein working toward an understanding of where sensor-based systems can best be leveraged. This work has several core aspects to be covered in this talk, including (1) starting from problems defined by municipal and community partners which can be translated into realistic technical specifications, (2) deploying sensor networks at a sufficient spatial density both to understand actual spatial variability created by the urban landscape and to determine spatial resolution needed to have confidence in resulting data products, (3) creating automated QA/QC tools to flag and categorize data related to sensor failures, anomalous weather, etc., (4) leveraging publicly available datasets and/or physics-based models in combination with sensor network datasets to meet stakeholder “data product” needs. Examples from multiple communities across Greater Boston will provide insight on the process and results to date.
    • Facilitating widespread action through Breathe London: A framework to support interpretation, comparison, and dissemination of high-quality air sensor data using visual media
    • Presentation by: Kayla Schulte, Environmental Research Group

      The global air pollution data and information landscape is expanding rapidly. Often data of varying quality is presented in a way that presumes knowledge of traditional, scientific formats (i.e. micrograms per cubic meter, AQIs). Evidence from the literature, alongside findings from Breathe London – the world’s first large scale hybrid monitoring network - reinforce the need for cogent data visualizations to support downstream actions. More specifically, visualizations should support “comparability” to help maximize information exchange and downstream actions for a variety of groups.

      To meet this demand, a four-step co-design framework was implemented to visualize hourly PM2.5 and NO2 data at over 400 locations on Version 2.0 of the Breathe London website. This process included background research, co-design workshops, evaluation, and website launch with feedback collection, while leveraging participatory methods,, the dialogue model, and social data science approaches.

      Target outreach groups, such as teachers, individuals with health conditions, decision-makers, and young people, were identified to ensure broad accessibility. Findings highlighted the importance of visual simplicity, customization, and features for comparing data from multiple sensors, and design to help facilitate action. Results also reinforced diversity of expertise within communities, exemplified by individuals creating their own data visualizations. This fed into adaptations to www.breathelondon.com, specifically: new visuals, instructional videos, and informational packets to enhance data interpretation and support downstream action.
    • Enhancing Air Quality Monitoring in the Paso del Norte Region Using Low-Cost Sensors
    • Presentation by: Leonardo Vazquez, Ph.D. Research Associate, The University of Texas at El Paso, United States

      The Paso del Norte (PdN) region on the US-Mexico border confronts significant health issues from particulate matter (PM) pollution. This study introduces enhanced air quality monitoring using low-cost PurpleAir sensors to analyze PM2.5 levels. A network of 26 monitoring sites was set up in El Paso, Texas, and Ciudad Juarez, Chihuahua, Mexico, offering real-time PM2.5 and meteorological data, fostering bi-national collaboration. Data accuracy was ensured through calibration and validation, revealing notable PM2.5 concentration disparities due to emissions and weather. Despite minor sensor issues, there was a strong correlation with FEM reference data (coefficients ~0.66-0.67). The network pinpointed PM2.5 hotspots, like ports and industrial areas, and refined data validation techniques. This work showcases the potential of affordable sensors in air quality monitoring for regulatory and research purposes, transforming data into actionable insights for air quality management and health research. It marks progress in air monitoring, aiding community engagement, policymaking, and cross-border cooperation.

    Towards a traceable quality of sensor data; from performance standards and test protocols to on-going data quality maintenance procedures

    Despite the common use of AQ sensors, the wider application of sensor data is hampered due to lack of (comparable) evaluation data and due to unknown or questionable performance . Performance standards and validation protocols for different pollutants are being developed to understand sensor performance under controlled and reproduceable conditions. In addition, there is a need to continuously assess sensor performance during their deployment under variable conditions and over time. Therefore, validated and traceable data quality procedures as well as maintenance and calibration protocols are needed.

    This session will highlight efforts to develop, compare, and validate performance targets and testing protocols on sensor systems, and on-going data quality control and maintenance procedures. It will address scientific insights on performance parameters, and on how to deal with sensor performance challenges over time. New insights in performance evaluation and on-going data quality procedures for air quality sensors are welcome!

    Session Chairs:

    Martine Van PoppelVITO, Edurne IbarrolaKUNAK

    Presentations: 
    • Experiences and evolution of sensor system QA/QC in the UK
    • Presentation by: Brian Stacey, Knowledge Leader, Air Quality Measurements, Ricardo, United Kingdom

      One of the key challenges for the use of low cost sensors is how to determine the quality of the data they provide.

      The process of standardization in Europe provides a start point for type testing and an indication of the likely level of data quality "out of the box", but many questions remain unanswered.
      This presentation will walk through the experiences in the UK, share some of the potential pitfalls and show how our manual QA/QC protocols have evolved to improve the quality of processed data.
    • Progress and advances in air quality sensors: from independent evaluations to standardization and certifications
    • Presentation by: Javier Fernandez, CEO, Kunak Technologies, Spain

      Over the past decade, there has been a worldwide effort to evaluate the performance of sensor-systems. Several independent evaluation efforts are occurring in Europe and the U.S. This is the case of the AQ-SPEC program and the AIRLAB Microsensor Challenge.

      Kunak has participated in all these evaluations, awarded as the most accurate multi-pollutant sensor in the Airlab Challenge 2021. These evaluations were the first attempt to show independent performance assessments of sensor systems. However, there were not enough for final users, showing the lack of an internationally accepted standard protocol that allows comparing the performance of instruments in different evaluation studies.

      In Europe, the CEN/TS 17660-1:2021 was developed, however, it is still under validation, while the USEPA created different evaluation protocols, but not an official certification. Thus, other certification bodies created certifications, such as MCERTS Certification in the UK, and INERIS in France.

      A lot of work has been done regarding sensor systems regulation, allowing consumers to understand the use of the devices, the performances, as well as their advantages and disadvantages. However, the work has not finished yet. Independent evaluations should have more precise methodology and protocols in terms of evaluation periods, ambient conditions and accuracy assessment. Besides, official regulatory standards must be aware that sensor systems will be a tool to complement the official regulatory network, with a technology lower in cost, so are not expected to perform as well as reference instruments. Thus, the requirements should be according to the technology and applications in which they are used. 
    • Evaluating Low-Cost Nephelometers: Challenges and Opportunities in High-Exposure Environments
    • Presentation by: Gulshan Kumar, Senior Research Scholar, Indian Institute of Technology, India

      Low-cost sensors (LCS) have the potential to provide spatiotemporal data for various applications, but their limitations include the absence of thorough evaluation data and uncertainties about their performance in widespread application scenarios.

      Our study, set in the context of South Asia, is a response to this pressing need. Within this region, numerous nations grapple with heightened PM2.5 concentrations, necessitating the creation of an extensive network of low-cost sensors. Given the region's dense population and elevated exposure levels, these low-cost sensors hold significant potential for bolstering community-driven scientific initiatives and non-policy applications. However, data quality concerns have often hindered their effectiveness.

      To address these concerns, we turn to the performance benchmarks set by the USEPA. It's noteworthy that these benchmarks have excelled in scenarios characterized by lower exposure levels and within a specific humidity range. To comprehensively evaluate sensor capabilities under more challenging circumstances, we conducted an extensive four-month outdoor study in Delhi. This study unfolded amid heightened exposure conditions (ranging from 6 to 611 micrograms per cubic meter, with an average of 134 micrograms per cubic meter) and elevated humidity levels (averaging 72%). Our findings indicate that while the sensors meet the guidelines at lower concentrations (<100 micrograms per cubic meter) and within the specific (40%-60%) humidity range, additional criteria are required for evaluating their performance at elevated levels. Furthermore, considering the diurnal variations in humidity, assessing the sensors with hourly performance metrics becomes essential for a temporal evaluation of LCS performance.

      In summary, our research aligns seamlessly with the conference theme, as it contributes to the ongoing efforts to establish performance targets, testing protocols, and data quality procedures for air quality sensors, with a particular focus on their deployment in challenging real-world conditions.

    • Lab Performance Evaluation of Lower-Cost Sensors for Smoke
    • Presentation by: Menaka Kumar, Data Analyst, National Student Services Contract, hosted by the U.S. Environmental Protection Agency, Office of Research and Development, United States

      The rise of lower-cost air sensors allows higher spatial and temporal resolved measurements of air pollutants, supplementing traditional regulatory air monitoring. Supplemental data is especially important during wildfire episodes when first responders and nearby communities are overwhelmed with high concentrations of smoke. Wildfire smoke is a complex mixture of air pollutants including fine particulate matter (PM2.5) and carbon monoxide (CO), which have well-known health impacts. Real-time PM2.5 and CO measurements near fires can be used to communicate with the public and incident command team to reduce smoke exposure. However, lower-cost sensors tend to be affected by atmospheric conditions and cross-sensitivities from interfering compounds causing some sensors to return inaccurate data, increasing uncertainty. This poor data quality results in the need to test sensor performance under elevated pollutant concentrations. Evaluations in elevated pollutant concentrations can be challenging to capture in the field due to infrequent smoke events but can be more easily achieved in an environmental test chamber. This work summarizes the results of laboratory evaluations of several lower-cost sensors such as the Lunar Outpost Canary X, the Lascar CO Data Logger, and the EPA developed Vehicle Add-on Mobile Monitoring System (VAMMS) relative to Federal Equivalent Method (FEM) instruments under high concentration conditions. In addition, the influence of environmental conditions, including temperature and relative humidity, are explored. This work adds to the on-going effort to improve lower-cost sensors’ data quality by providing a better estimate of accuracy. Also, this work is the first evaluation of these carbon monoxide (CO) sensors for wildfire smoke applications. As a considerable number of research and community organizations are interested in utilizing lower-cost sensors, evaluating their performance in smoke conditions in a controlled environmental test chamber allows for greater data confidence, greater understanding of local air quality trends, policy change, and informed sensor selections. 

      Although this abstract was reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

    • US EPA Performance Testing Protocols, Metrics, and Targets for Air Sensors
    • Presentation by: Andrea Clements, Physical Scientist, US EPA, United States

      As the development and use of air sensors continues to expand, understanding the performance of sensors remains a critical need given the variability in sensor data quality. To help support consumers and developers of air sensor technologies, the U.S. Environmental Protection Agency (U.S. EPA) published reports in 2021 and 2023 outlining recommended testing protocols, metrics, and target values to evaluate the performance of air sensors used in non-regulatory supplemental and informational monitoring applications (NSIM). The 2021 reports covered ozone (O3) and fine particulate matter (PM2.5) sensors, and the 2023 reports were supplemental reports discussing sensors measuring particles with diameters of 10 microns or less (PM10), nitrogen dioxide (NO2), carbon monoxide (CO), and sulfur dioxide (SO2). The goals of these efforts are to provide a consistent approach for evaluating sensor performance, provide guidance on appropriate testing environments (where applicable), and encourage consistent reporting of sensor evaluation results. Additionally, the reports aim to provide confidence in sensor data quality, help guide future technology improvements, and assist users in selecting the appropriate sensors for their NSIM application of interest. This presentation will summarize the U.S. EPA’s work to date in developing guidance for evaluating the performance of air sensors.

      Disclaimer: Although this abstract was reviewed by EPA and approved for presentation, it may not necessarily reflect official Agency policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

    • Towards traceability of low-cost gas sensors: new approaches to sensor evaluation, validation and certification
    • Presentation by: Valerio Ferracci, Senior Scientist, National Physical Laboratory, United Kingdom

      Low-cost sensors have the potential to achieve real-time, hyper-local air quality monitoring, and therefore to complement traditional reference instrumentation in providing more detailed information on air pollution. However many aspects of their performance relative to reference methods remain to be assessed, and approaches based on metrology can help validate performance targets and testing protocols for these novel sensor systems.

      Here we present a two-pronged approach towards more robust, traceable measurements from low-cost sensors. Firstly, we developed the Multiple Atmosphere Controlled Environment (MACE), a state-of-the-art type-testing facility to traceably demonstrate the ability of low-cost sensor systems to meet performance requirements, such as those in the recent Technical Specification CEN/TS 17660-1:2021. MACE can generate atmospheres with trace gas pollutants such as NO, NO2, O3, CO, SO2, CO2 and NH3 under controlled environmental conditions (e.g., temperature, relative humidity). Secondly, NPL led the development of a Publicly Available Specification (PAS 4023) to provide recommendations for the selection, deployment and quality assurance of low-cost air quality sensor systems and aimed at a variety of users, from local authorities to citizen scientists. This presentation discusses highlights from these two approaches in improving data quality and comparability for low-cost sensors.

    • ASTM Standards and Specifications for Outdoor Air Quality Sensors
    • Presentation by: Geoff Henshaw, Aeroqual

      Over the past fifteen years, the availability of lower priced air quality sensors has encouraged innovative measurement of air quality by a wide range of users. Over the past few years organisations in the EU, US, China and elsewhere have developed standards and methods to objectively evaluate air sensors and create criteria for performance-based grades. This presentation provides an update on the work undertaken by ASTM International, a consensus based standards organization.
      ASTM established a working group in 2018 to develop standards for the evaluation and specification of outdoor air sensors focused on the criteria pollutants: CO, SO2, NO2, O3, PM2.5 and PM10. The ASTM air sensor evaluation standard D8406 was published in 2022. It includes both laboratory and field tests that provide information on instrument repeatability, sensitivity, linearity, cross- interferences, drift, and comparability with Federal Equivalent Methods.
      ASTM began work in 2020 on a specification for outdoor air sensors that would create grades of performance based on the standard test practice. The key criteria and their rationale will be presented and compared with other standards. The steps to approval and publication will discussed.
      New standards work by ASTM on non-regulatory air quality monitoring will also be described.

    Leveraging low-cost air quality sensors for exposure assessment and health effects research

    Low-cost air quality sensors provide an accessible and affordable way to assess air pollution exposures, which have implications for health outcomes. Sensors provide opportunities for high spatiotemporal resolution measurement. However, they also present data quality, maintenance, and interpretation challenges. Speakers in this session will share experience and discuss the use of sensors in exposure assessment and health research, including strategies for use, benefits, and challenges.

    Session Chairs:

    Kofi AmegahUniversity of Cape Coast, Orly StampferWashington State Department of Health, Yisi Liu, USC, Liam O'FallonNIEHS

    Presentations:
    • Measuring air quality impacts and exposure during prescribed burning activities in Colorado
    • Presentation by: Evan Coffey, Sr. PRA, University of Colorado Boulder, United States

      Wildfire impacts are growing in the US and around the world due to climate change, fire suppression policies, and growth and expansion of the wildland-urban interface (WUI). While prescribed (Rx) burning can reduce wildfire impacts, it also entails its own risks, including the risk of “escape” and risks associated with exposure to smoke for nearby WUI communities and burn implementers. Our team of interdisciplinary researchers have been collaborating with federal, state and municipal Rx burn implementers to deploy instruments including custom-built air quality monitors, called “pods”, to measure particulates and gaseous pollutants emitted from Rx fires. In addition to the pods placed directly near the burn units, pods are positioned in several rooms inside of homes of participating households as well as outdoors to monitor indoor and outdoor air quality and estimate exposure impacts from the Rx fire. Air quality measurements, surveys and health metrics taken before, during and after each burn event are being investigated to better understand how Rx burning activities can impact air quality, health, and public perceptions. Challenges to making these measurements and preliminary results from the first half of this 3-year project will be presented.
    • Combining PM sensor data and particle types measured by passive samplers to evaluate affordable clean air solutions in farmworker communities (FRESSCA project)
    • Presentation by: Jeff Wagner, Chief, Environmental Health Laboratory Branch, California Department of Public Health, United States

      Results are presented from combined PM sensor data and particle types measured by low-cost, passive samplers to evaluate affordable clean air solutions in farmworker communities. Agricultural workers in California’s San Joaquin Valley are disproportionately exposed to air pollution. Many farmworkers have evaporative ("swamp") coolers (ECs) in their homes. ECs bring in unfiltered outdoor air, creating an indoor hazard when outdoor pollution episodes and heat coincide. The Filtration for Respiratory Exposure to wildfire Smoke in Swamp Cooler Air (FRESSCA) project employed a community-based, multidisciplinary team to design DIY filters for ECs and test them in agricultural workers’ homes. Portable air cleaners were provided to participants as well. These clean air interventions were tested using continuous indoor/outdoor PM2.5 monitoring and passive PM samplers, as well as simultaneous chemical, environmental, and biomonitoring measurements. The passive PM samplers were designed to be field deployable and retrievable by community staff, then analyzed in the lab using electron microscopy to determine the major particle types, e.g., windblown dust, combustion PM, and wildfire smoke particles, both outdoors and indoors with different filtration solutions. The passive PM samplers also provided size distributions, PM2.5, and coarse PM. The results from this project demonstrate the use of leveraged low-cost data to improve health equity and address combined air pollution and thermal stress in a changing climate. 
    • Assessment of NO2 and PM Exposure of Asthma and COPD Patients Using Sensors 
    • Presentation by: Miriam Chacón Mateos, Research Scientist, University of Stuttgart, Germany

      The use of air quality sensors for environmental epidemiology has many advantages, for instance, the decrease in the bias of exposure estimations when compared with fixed outdoor monitoring stations.


      A pilot study using PM and NO2 sensors was carried out in Stuttgart (Germany) in cooperation with University Hospital Charité in Berlin and a pneumonology Praxis in Stuttgart. For this purpose, two sensor systems were designed for indoor and outdoor measurements. After laboratory and field calibration, the sensor systems were deployed in the houses of seven patients with Asthma or COPD (chronic obstructive pulmonary diseases) during 30 days. Participants completed spirometry, a questionnaire assessing respiratory symptoms and a protocol of activities at a daily basis. They also filled a questionnaire about the house conditions at the beginning of the campaign and a feedback questionnaire at the end. In this talk, we will show the results of the source apportionment using the information from the protocol of activities and the sensor data, the preliminary associations between exposure to NO2 and PM2.5 and acute health outcomes and we will report about the feedback received from the patients and lessons learned including data quality assurance. 
    • Quantifying long-term exposures to fine particulate matter (PM2.5) with real-time, low-cost sensors to assess the impact of household air pollution on birthweight in the TAPHE-II cohort, India. 
    • Presentation by: Naveen Puttaswamy, Assistant Professor, Sri Ramachandra Institute of Higher Education and Research, India

      Exposures to PM2.5, typically measured for 24 or 48-h in health effects research, lack temporal resolution. Real-time, low-cost sensors (LCS) offer potential long-term PM monitoring solutions with high spatio-temporal resolution needed for exposure – response analyses. We aimed to assess long-term exposures to PM2.5 among pregnant women and its effect on birthweight in the TAPHE-II cohort using real-time LCS. We used real-time PM sensors equipped with PMS7003™ (Plantower Inc., China) to monitor living room PM2.5 levels for a period of 21 days in 80 homes. Sensors were collocated with gravimetric samplers to develop indoor-specific calibration equations. In addition, all sensors were collocated with a reference-grade monitor (i.e., BAM1022); linear models were fit to derive calibration coefficients. Continuous PM data was collected on average (s.d.) for 21 (3) days; data availability ranged between 97 to 100% across all homes. Precision across all sensors was satisfactory with a standard deviation of 2.6 µg/m3 and a coefficient of variation of 15.6%. The NRMSE for indoor and ambient collocation was 31.6% (r=0.86) and 43.2% (r=0.80), respectively. Correlation (NRMSE) between personal 24-h and 21-day LCS PM2.5 measures was 0.62 (47.9%). In adjusted linear models, an interquartile increase in exposure to PM2.5 was associated with 83.2 g (95% CI: 27.4 – 138.0, p=0.004) reduction in birthweight. We show the applicability of LCS for long-term indoor PM monitoring to assess health risks associated with HAP. 
    • The Impact of Indoor PM2.5 Exposures on Sleep Quality: Insights from Community Based Low-Cost Sensing 
    • Presentation by: Tianjun Lu, Assistant Professor, University of Kentucky; California State University, Dominguez Hills, United States

      Air pollution exposure may adversely impact sleep health, yet the relationship between indoor air quality and sleep quality remains understudied. We used low-cost PurpleAir monitors to continuously measure indoor PM2.5 data over 22,836 hours from 11 households across disadvantaged communities in South Los Angeles from June to September 2022. We administered five bi-weekly surveys per household, collecting information on sleep quality (e.g., time taken to fall asleep, bedtime, and thermal discomfort perceptions). We developed mixed-effect linear models to investigate the association between hourly indoor PM2.5 concentrations and time taken to fall asleep after adjusting for temperature, relative humidity, and day of week. We tested the impact of daily PM2.5 concentrations on time taken to sleep, and the lag effect of exposures during the 1-6 hours before bedtime. On average, the hourly PM2.5 levels peaked on Tuesdays in a week (mean=14.9 µg/m3, SD=20.6 µg/m3) and between 8:00 am and 12:00 pm in each day (mean ranged from 14.4 to 17.4 µg/m3, SD ranged from 15.3 to 22.6 µg/m3). For each 10µg/m3 increase in 24-hour indoor PM2.5 concentrations, the time taken to fall asleep increased by 0.08 min (95% CI: 0.01, 0.14). Notably, we found that PM2.5 concentrations around 3 hours before bedtime held the greatest impact, that time taken to fall asleep increased 0.22 min (95% CI: 0.03, 0.41) for each 10µg/m3 increase in hourly indoor PM2.5. These results underscore the influence of indoor PM2.5 on sleep quality. Future interventions may target indoor air quality 3 hours before bedtime. 
    • Invited Presentation by: Jose Guillermo Cedeño Laurent, Rutgers University
    • Presentation title and description forthcoming.
    • Invited Presentation by: Sigride Asseko, Oregon State University
    • Presentation title and description forthcoming.

    Indoor air: Sensors Driving Indoor Air Optimization

    The COVID-19 pandemic and the ongoing events of smoke from forest fires have brought to fore concerns around indoor air and the role of ventilation, filtration, humidity control, etc in addressing it.  The pandemic raised questions around the adequacy of existing guidelines to contain infections indoors and the need for increased outdoor air fraction in the supply line.  With the forest fires, however, concerns shifted to limiting outdoor air and filtration.  These competing requirements, coupled with demands to improve energy efficiency in buildings, create an opportunity for novel sensing based solutions for control of ventilation systems.  In this track, we invite presentations focused on the role of sensors in indoor air monitoring and control.

    Session Chairs:

    Suresh DhaniyalaClarkson University, Randy ChapmanUS EPA

    Presenters: 
    • Impact of Wildfire Smoke on Indoor Air Quality Using Low-Cost Sensors.
    • Presentation by: Mst Rowshon Afroz, PhD Student, Environmental and Atmospheric Chemistry, Department of Chemistry, University of Alberta, Canada

      Wildfires have become more frequent and severe in Western Canada in recent years, which has had a negative impact on air quality. Although homes are considered a barrier against outdoor pollutants, wildfire contaminants can still infiltrate indoor spaces. Even though the harmful effects of wildfires are well-known, the extent to which they impact indoor air quality (IAQ) is still uncertain. Since people spend most of their time indoors (about 90%), monitoring IAQ is crucial for their health and safety. In recent years, low-cost sensors (LCSs) have become popular compared to traditional measurement approaches for real-time monitoring of IAQ in diverse building settings due to their affordability, portability, user-friendliness, and ease of operation. Our project aims to investigate how wildfire smoke affects IAQ using the Internet of Things-based LCS. Specifically, we want to determine the indoor-to-outdoor (I/O) ratios of PM2.5 in various buildings and explore any possible correlation between the building's purpose, ventilation schemes, and its I/O ratios. In addition to PM, our LCSs also monitored CO2, temperature, and relative humidity. Between May to September 2023, we deployed 33 LCS units across various buildings on the University of Alberta (UofA) campus through community engagement. These units were installed in laboratories, libraries, office buildings, residences, gyms, hospitals, and other structures.  UofA's wide variety of structures could provide valuable insights into the effects of wildfire-generated pollutants throughout Canadian residences. During the study period, real-time sensor data was integrated into a website, providing the campus community with information on their building's air quality. Our research findings will suggest strategies for reducing smoke particle infiltration in buildings through sustainable building practices and will also help mitigate the risks of exposure to wildfire pollutants.
    • Integrating Low-Cost Sensor Networks for Enhanced Indoor Air Quality Monitoring and Optimization of Ventilation Systems 
    • Presentation by: Chethani Athukorala, Graduate Student, Clarkson University, United States

      Traditional indoor ventilation systems recirculate air from the space, mix it with outdoor air, and then reintroduce it into the environment following filtration. These systems largely only control thermal comfort, ignoring indoor air quality. The recent pandemic and wildfire smoke events has brought urgency to monitor indoor air and integrate that with the operation of ventilation systems. The supply of safe indoor air must be made while ensuring sustainable energy use in the built environment. This is a tricky balance and our efforts to achieve this must be validated with real-world use case. 


      In our study, we chose a heavily used University lecture room for real-time monitoring of thermal comfort and air quality using low-cost networked sensors over a two year period and quantified the health risks and energy usage of the space under the existing ventilation system operation. The sensor data was then correlated with HVAC operating conditions, space utilization patterns, and outdoor environmental factors. Subsequently, we integrated indoor air data with HVAC operating conditions to develop a machine learning model to establish a relationship between ventilation system operation and infection risk for space users.

      The resulting model was employed to optimize HVAC operation, ensuring the maintenance of indoor air safety while minimizing energy consumption. This innovative approach provides insights into enhancing air quality management strategies and presents a practical framework for optimizing ventilation systems in real-world indoor environments.
    • Characterization of Indoor and Outdoor Concentrations of PM2.5, ultrafine particles, and black carbon at homes in Dhaka, Bangladesh
    • Presentation by: Aynul Bari, Assistant Professor, University at Albany, State University of New York, United States

      Bangladesh is a global hotspot for particulate air pollution, resulting in significant premature deaths and economic losses. While there is limited outdoor monitoring data for total PM2.5 concentrations in Bangladesh, there is a notable absence of characterization for indoor exposure levels. As part of NSF-IRES (International Research Experiences for Students) projects, this study aims to address these gaps in monitoring data by characterizing indoor PM2.5 levels, including size and chemically specific PM2.5 markers such as ultrafine particles and black carbon. We deployed PurpleAir PM2.5 sensors both indoors and outdoors in ten homes across Dhaka City, collecting continuous data over a 2-year period. We also measured ultrafine particle number concentration using a condensation particle counter (CPC) and black carbon using ObservAir detector (filter-based light absorption at 880 nm). These measurements were conducted during short-term visits (30 minutes/visit) at selected homes, with visits occurring on 7-10 different days in both wet and dry seasons. We also collected information about physical features and indoor sources within each home through a structured questionnaire survey. Our findings will provide an improved understanding of community-specific particulate air pollution levels, increase public awareness, and inform policy makers the need to reduce indoor and near-field outdoor exposure.
    • Fusing low-cost sensor data, occupant feedback, and energy use for optimal IAQ in residential settings
    • Presentation by: Dennis Heidner

      Indoor Air Quality (IAQ) is a subset of the larger Indoor Environment Quality (IEQ) in buildings.  In this presentation will focus on our work developing low cost IoT air quality sensing devices that pair up with occupant-perception survey devices.  

      Using multiple low-cost sensors in a single package we collect a large dataset in a residence.  Low-cost sensing devices allow measurements to be taken continuously at many  places within a building and to build redundancy into the data. A robust dataset can contain samples collected over time and for a wide range of homes and occupants.  The data collected is then used to train a “TinyML” AI device which could optimize the operation of the HVAC system for the residence – optimizing the IAQ, as well as the IEQ, and energy use.

      Just learning the data from the sensor is not enough to ensure a safe IAQ.  Occupants in the residence will not detect the harmful levels of CO, CH4 or some VOCs.  Frequently the health impacts are dependent of the exposure times; multiple exposure/risk scenarios such as average daily exposure, or maximum single-time exposure, can be trained into the AI network as constraints. Combined temporal perception, sensor data, exposure risks, building science and energy costs can be used to teach an AI system about optimal IAQ/IEQ.

      We also present the idea of radar charts overlaying what’s been sensed, occupant feedback, and mitigation strategies. These charts allow both occupants and professionals to visualize the results in a wholistic manner.

      This presentation focuses on the devices and methods we see as important for creating the low cost IoT sensors assemblies.  Our effort has been to develop open source/hardware devices and training material for high school and college undergraduate levels, thus helping future generations to think sensors, IAQ/IEQ, and building science for homes.
    • Characterizing indoor air quality improvements associated with electric heating, cooking, and smart filtration appliances in disadvantaged communities in the San Joaquin Valley: Mid-Study Results
    • Presentation by: Katie Kearns, Technical Director, Berkeley Air Monitoring Group, United States

      California’s San Joaquin Valley is in the 90th percentile for pediatric asthma statewide, as well as environmental exposure risk from PM2.5 air pollution. Recent studies and corresponding media attention have raised concerns, particularly regarding childhood asthma, over exposures to gas stove emissions in homes. This ongoing study is investigating air quality impacts following a gas-to-electric stove transition by leveraging the California Public Utilities Commission (CPUC) program providing free induction stove upgrades to low-income rural households. Homes are being selected among eligible CPUC program participants (target n=75 with electric heating and cooking), matched by a group of homes using biomass and propane appliances (target n=75). Each home is monitored monthly for pollutants including PM2.5, NO2, NOx, and CO. Data from the first 84 homes indicate significantly lower NO2 levels in electric (1.6 ppb) compared to gas stove homes (12.9 ppb). NOx levels mirrored this trend. CO levels were low across the board, but with notable high outliers in gas stove homes. PM2.5 levels showed minimal differences, aligning with expectations as gas stoves emit negligible PM2.5. These findings suggest that induction stoves contribute to reduced indoor NO2 and NOx levels.
    • Relative contributions of ambient air and internal sources to multiple air pollutants in public transportation modes
    • Presentation by: Zhiyuan Li, Associate Professor, Sun Yat-sen University, China

      Commuters are often exposed to relatively high air pollutant concentrations in public transport microenvironments (TMEs) because of their proximity to emission sources. Previous studies have mainly focused on assessing the concentrations of air pollutants in TMEs, but few studies have distinguished between the contributions of ambient air and internal sources to the exposure of commuters to air pollutants. The main objective of this study was to quantify the contributions of ambient air and internal sources to the measured particulate matter and gaseous pollutant concentrations in selected TMEs in Hong Kong, a high-rise, high-density city in Asia. A sampling campaign was conducted to measure air pollutant concentrations in TMEs in Hong Kong in July and November 2018 using portable air quality monitors. We measured the concentrations of each pollutant in different TMEs and quantified the infiltration of particulate matter into these TMEs. The double-decker bus had the lowest particulate matter concentrations (mean PM1, PM2.5, and PM10 concentrations of 5.1, 9.5, and 13 μg/m3, respectively), but higher concentrations of CO (0.9 ppm), NO (422 ppb), and NO2 (100 ppb). For all the TMEs, about half of the PM2.5 were PM1 particles. The Mass Transit Railway (MTR) subway system had a PM2.5/PM10 ratio of about 0.90, whereas the PM2.5/PM10 ratio was about 0.60–0.70 for the other TMEs. The MTR had infiltration factor estimates <0.4 for particulate matter, lower than those of the double-decker bus and minibus. The MTR had the highest contribution from internal sources (mean PM1, PM2.5, and PM10 concentrations of 4.6, 13.4, and 15.8 μg/m3, respectively). This study will help citizens to plan commuting routes to reduce their exposure to air pollution and help policy-makers to prioritize effective exposure reduction strategies.
    • Understanding the effect of outdoor pollution episodes and HVAC settings on indoor air quality using networked sensors
    • Presentation by: Tristalee Mangin, Research Assistant, University of Utah Chemical Engineering, United States

      Individuals spend up to 90% of their time indoors, and poor indoor air quality (IAQ) is associated with numerous adverse human health effects. Yet, it has been less frequently studied than outdoor air quality.  Particulate matter (PM), specifically PM2.5 concentrations, is a key driver of adverse health effects.  This study aims to understand the impact of pollution episodes (wildfire smoke, dust events, and temperature inversions) on indoor air quality using a network of low-cost air quality sensors on a university campus and to identify the effect of HVAC operations on IAQ.  

      This study deployed 21 low-cost air quality sensor nodes measuring PM2.5 and carbon dioxide (CO2) levels, with 17 sensors at indoor locations and four sensors at two outdoor locations across the University of Utah campus. We developed a dashboard to share the AQ measurements in real time with facilities management.  The sensors were deployed in November 2022 and continue to operate.  We identified the following outdoor pollution events: four temperature inversion events (when the valley heat deficit values exceeded 4.04 MJ/m2 for three or more days), 10 dust events (when the regulatory PM10 concentrations exceeded 100 ug/m3 with a wind speed greater than 5 m/s), and two wildfire smoke events (using the aerosol optical depth results from NASA WorldView satellite images).  The University of Utah facilities management has provided access to the HVAC settings and equipment measurements for these pollution episodes.  

      Wildfire smoke has the largest impact on IAQ of the three pollution events studied.  These events had the highest indoor-to-outdoor PM2.5 correlation with an R2 value of 0.68, while inversion and dust events had indoor-to-outdoor correlation values of 0.32 and 0.44, respectively.  The wildfire smoke events resulted in the highest hourly average PM2.5 concentrations for 12 of the 17 indoor locations compared to inversion and dust events. During the wildfire events, 9 locations exceeded World Health Organization 24-hour PM2.5 concentration guidelines for a total of 13 days, with a maximum hourly PM2.5 concentration of 37.4 ug/m3. This is a surprising finding, given that many of the buildings are equipped with MERV13 filters, which have an efficiency rating of >50% for PM ranging from 0.3 – 1 um and >85% for PM ranging from 1 – 3 um, both of which are associated with wildfire events. Preliminary HVAC analysis indicates that HVAC operations impact IAQ during pollution events.  For example, the team found one location where an open outdoor damper resulted in an increase in the indoor-to-outdoor PM2.5 concentration ratio from 0.33 to 0.79.  Further analysis of HVAC operations is underway for the additional locations and different pollution events. 
    • Invited Presentation by: Theresa Pistochini, UC Davis
    • Presentation title and description forthcoming.
    • Distributions of PM2.5 in indoor air of Households during cooking and daily activities: A Pilot Study
    • Presentation by: Mahesh Senarathna, Research Assistant, Postgraduate Institute of Science, University of Peradeniya, Sri Lanka

      Particulate matter (PM) concentration in indoor air significantly determines human health and well-being in households. Long-term PM2.5 exposure is known to raise the risk of developing serious health problems. Households using traditional biomass stoves for cooking can contribute to high levels of PM2.5 indoors and outdoors; however, the level of PM2.5 content in households in Sri Lanka is not known. This study aimed to measure PM2.5 levels in indoor environments in Sri Lanka. PM2.5 concentrations were continuously monitored using calibrated low-cost air quality sensors for one week in three households where traditional daily living activities take place. The study deployed small sensors in living room, dining room, bedrooms, and kitchen, and one sensor outside the house after obtaining informed consent from the chief occupant.  An activity diary was maintained to record the different types of burning performed for the entire week. Particulate matter data were analyzed in conjunction with the burning activity diary.  There was a statistically significant difference in PM2.5 concentration measured at various activity times as determined by one-way ANOVA (F = 17.28, p < 0.05). The highest levels of PM2.5 were reported when multiple burning activities occurred simultaneously, such as cooking with biomass fuel and burning incense powder and incense sticks. The average peak concentration reached up to 385 ± 450.67 µgm-3 in the living room and 342.88 ± 365.67 µgm-3 in the kitchen during this time. Furthermore, the Kitchen was found to have the overall highest average PM2.5 level (81.97 µgm-3), while just outside the house shows the lowest (26.84 µgm-3). The levels of PM2.5 were increased in every part of the house when the burning activities were happening inside the household. The findings emphasize the critical importance of comprehending the dynamic nature of indoor air quality and its relationship to household activities. Recommendations for effective mitigation strategies include better ventilation and targeted interventions during high PM2.5-producing activities. This research can help policymakers, building designers, and residents improve indoor air quality and reduce potential health risks associated with elevated PM2.5 levels.
    • Invited Presentation by: Brett Singer, LBL
    • Presentation title and description forthcoming.
    • Air Monitoring in Mechanically Ventilated Classrooms Using Low-cost Sensors
    • Presentation by: Zhong-Min Wang, Research Scientist Supervisor,  Lab Unit Chief, California Department of Public Health, United States 

      A pilot field monitoring study of air quality in four mechanically ventilated classrooms was conducted in Northern California using low-cost sensors.  The purpose of this study is to investigate particulate matter (PM) under different ventilation and filtration conditions to improve the air quality in California schools. In this study, we measured both indoor and outdoor PM concentrations and compositions during a 7-week period in 2022 in occupied classrooms using two different low-cost sensors and open-face passive aerosol samplers. We also monitored the CO2 concentration in each classroom studied to determine the relationship of indoor CO2 concentration with ventilation and student activities. The results showed that HVAC MERV 13 filters played an important role in improving indoor air quality, especially for PM2.5 reduction. The passive aerosol samplers deployed in this study provided a direct observational comparison of particle morphology differences between indoor and outdoor air. This work also suggested that the installation and application of low-cost sensors in schools had a positive impact on improving air quality and increasing students’ environmental awareness. The results also provided evidence that window and door opening behavior could impact indoor air quality, which may be especially relevant during bad outdoor air quality days such as the wildfire episodes.
    • An Application of Low-cost air quality sensors in indoor poultry facilities 
    • Presentation by: Ran Zhao, Assistant Professor, University of Alberta, Canada

      Low-cost sensors (LCS) can offer affordable indoor air quality monitoring for industries facing unique and elevated indoor air pollutants. In Canada, due to the cold winter, poultry farms are entirely indoors. Concentrations of particulate matter (PM) and other air pollutants can reach concentrations orders of magnitude greater than outside, putting the occupational health of producers and the welfare of birds at risk. Research-grade instruments are not suitable for long-term monitoring there due to the cost of instruments and the dusty environment. In this work, my team aimed to develop and apply an LCS for indoor poultry facilities. The measurement took place at a commercial table egg farm in Alberta, Canada during winter, when the ventilation rate of barns was reduced to combat cold temperatures outdoors. After many trials, we assembled an LCS that could continuously monitor PM, CO2, temperature, and RH for up to months in the farm. By applying a single correction factor, we were able to bring the sensor PM data into a reasonable agreement with a reference Optical Particle Counter. Further, the LCS provided insights into the trend and correlations of air pollutants in the farm, which was achievable only with research-grade instruments previously.  

    Next Generation Sensing

    Current air quality sensors focus on criteria air pollutants and have well-known limitations. We invite talks on sensors for emerging and non-regulated pollutants, new technology to measure existing pollutants, and other sensor hardware innovations. Talks may also explore what future sensor developments are needed.

    Session Chairs:

    R. SubramanianCSTEP, Karoline BarkjohnUS EPA

    Presentations: 
    • Low-cost Methods for Measurement of PM2.5 Composition at African cities by Exploiting Existing Beta Attenuation Monitors 
    • Presentation by: Abhishek Anand, PhD Student, Carnegie Mellon University, United States

      A lack of continuous air pollution monitoring due to high cost of regulatory monitors and severely high health burden creates an urgent need for cost-effective methods to measure air pollutants in the Global South. We introduce a low-cost technique to quantify black carbon (BC) by utilizing reflected red light from particle deposits on filters. This method uses filter tapes from already existing PM2.5 monitors, known as Beta Attenuation Monitors to estimate hourly BC concentrations. An hourly PM2.5 deposit (spot) is photographed on a reference card. The photo goes through a custom Computer Vision-based algorithm for computing red channel value (R) for the spot and using R to get the BC level for that hour with a R-to-BC calibration model. The method demonstrates an effective detection limit of ~0.15 µg.m-3 for 1-hr samples. We present 2-month hourly BC measurements and diurnal patterns for four cities in Africa, including Abidjan, Addis Ababa, Accra, and Algiers. The average BC between July-August 2020 in Abidjan was 3.85 µg.m-3, ~10 times higher than that in Pittsburgh (USA). The average BC/PM2.5 ratio during this period was found to be ~18%. The BC data, thus obtained, can be crucial in identifying emission sources and effective policymaking.
    • RADICAL: Developing an electronic sensor for detecting short-lived atmospheric radicals and other gases
    • Presentation by: John Wenger, Professor of Chemistry, University College Cork, Ireland

      Atmospheric radicals, particularly hydroxyl and nitrate, are the drivers of chemical processes that determine atmospheric composition and influence air quality and climate. However, the detection of these short-lived atmospheric radicals is far from routine, and only a few labs worldwide can accurately measure their concentrations in air. Current techniques for measuring radicals are based on spectroscopic and mass spectrometric methods, which although sensitive and robust, are technically complex, cumbersome and expensive.


      This presentation provides an overview, and a discussion of the latest results, from the EU-funded project ‘RADICAL’ which is developing a small, affordable sensor to electrically detect short-lived atmospheric radicals in real-time. This will be the first gas sensor built from an array of junctionless nanowire transistors, which have proven popular for liquid-based sensors. These silicon devices are functionalised with organic “probe” molecules (e.g. alkenes) for high selectivity and sensitivity towards the target analytes. Preliminary results for these sensors show ppb-level detection capability towards different atmospheric species including NO2, O3 and HOx. 

      Although challenging, RADICAL sensors not only have the potential to be deployed on a large scale but can also be adapted to detect other important atmospheric gases, particularly on short-timescales. The project team welcomes ideas for future collaborations on how these sensors might be best applied in real-life environmental monitoring situations.
    • Using the Figaro Taguchi Gas Sensor to measure methane mole fraction with a baseline reference resistance 
    • Presentation by: Adil Shah, Postdoctoral Research Scientist, Laboratoire des Sciences du Climat et de l'Environnement, France

      Cheap and widespread in-situ methane mole fraction ([CH4]) measurements are required to detect and quantify methane emissions and hence, to help to constrain the global methane budget. The low-cost semiconductor-based Figaro Tagchi Gas Sensor (TGS) can detect reducing gases, including methane. But converting measured TGS resistance into [CH4] is not straightforward. This is further complicated as the TGS is sensitive to environmental variables including temperature and water mole fraction ([H2O]). To overcome this, a reference resistance can be modelled at a reference [CH4] level, as a function of (non-gas) environmental conditions. The ratio between measured TGS resistance and modelled reference resistance yields an excellent two-term modified power fit as a function of [CH4]. The TGS 2611-C00 was therefore used to derive [CH4] from a landfill site. A reference resistance was derived in the field as a function of temperature and [H2O] from background sampling, identified using wind direction measurements. Resistance ratio could then be used to yield [CH4] with an accuracy of better than ±1 ppm.
    • A Deep Dive into a Next Generation BTEX Ambient Air Monitor 
    • Presentation by: Heather McIntyre, Physical Sciences Researcher/Scientist II, CDPHE APCD ATOPS OMM, United States

      Abstract: Benzene, toluene, ethylbenzene, and xylenes (BTEX) are an important subset of air toxics compounds in ambient air. Benzene is emitted from a variety of sources and is one of the primary health risk drivers found in ambient air due to its toxicity. BTEX is typically monitored in air using gas chromatography, which is accurate and reliable, but very resource intensive. With advancements in technology, several new analytical techniques are available to monitor BTEX that require fewer resources, less infrastructure, and are simpler to operate. Micro-GCs are a next generation emission monitoring (NGEM) technology that shows promise in being utilized for long-term, stationary ambient BTEX measurements for air quality agencies across the country.


      To investigate the applicability of utilizing these new technologies as viable community monitoring analyzers, the Colorado Department of Public Health and Environment (CDPHE) conducted a co-location study with a Pyxis Micro-GC and a MARKES-Thermo Fisher benchtop GC  near a new oil and gas development site. This presentation will explore and discuss the level of agreement observed between the two units, the strengths and weaknesses of each technique, and next steps needed for advancing the usage of the Pyxis Micro-GC technology in air quality networks nationwide.
    • Highly stable low-cost Ag/MXene sensors for formaldehyde detection 
    • Presentation by: Shwetha Sunil Kumar, PhD Student, Carnegie Mellon University, United States

      Formaldehyde, a potential carcinogen, is one of the most common volatile organic compounds present in air due to its widespread indoor and outdoor sources. Currently, the state-of-the-art instruments for formaldehyde detection are expensive, require trained operators or are too bulky to be deployed in a well-distributed spatial network. This study aims to develop low-cost, portable, and reliable chemiresistive sensors for the detection and quantification of formaldehyde. We use Ti3C2Tx, which belongs to a class of layered materials called MXenes, as the sensing element. MXenes are two-dimensional transition metal carbides and nitrides, which have recently come to the fore as potential sensing candidates specifically due to their high conductivity, tunable surface chemistry and 2D structure. In this study, Ti3C2Tx is first doped with nitrogen to increase its d-spacing and then used to form a p-n heterojunction with silver nanoparticles (Ag NPs). This hybrid formation leads to an increase in the sensor response. A major barrier to the commercial use of MXene sensors is their susceptibility to oxidation, which results in loss of conductivity and ultimately performance degradation. To overcome this, we encapsulated these sensors. The encapsulation was found to drastically improve the stability of the sensor, with a degradation of less than 4% over 2 months compared to a 100% degradation in the unencapsulated case over the same time period. The extent of N-doping, concentration of Ag NPs and thickness of the encapsulant layer were optimized to reach a limit of detection (LOD) of 0.9 ppm. Reproducibility and selectivity tests for these encapsulated Ag/MXene sensors were also carried out.
    • A Condensation Particle Counter for Community Monitoring 
    • Presentation by: Susanne Hering, Senior Research Scientist, Aerosol Dynamics Inc., United States

      With the increasingly widespread monitoring of air pollutants at the community level, there is a growing need for accurate and reliable data at a reasonable cost. Especially challenging are measurements of ultrafine or nanometer sized particles, as these are too small to be detected directly by optical sensors. Reported here is a new condensation particle counter that targets this community measurement need. It utilizes quasi-adiabatic expansion coupled to single particle counting to monitor particle number concentrations in the 5-2500 nm size range. The expansion of humidified air sample creates a region of supersaturation leading to the condensational growth of the particles within the cell. The droplets formed are counted individually as they exit the cell, while the sample volume is determined from the concurrent change in pressure in the chamber. This yields a 'first principles' measurement of ultrafine particle concentrations. The counting efficiency is 50% at 5nm; above 80% at 10nm, and above 95% for particles larger than 30nm. Ambient monitoring alongside a research grade condensation particle counter gives with a 5nm threshold yield regression slope above 0.9, and a correlation coefficient squared of 0.98.  
    • Performance Evaluation Results for the Inaugural Set of VOC Air Quality Sensors Tested Under the AQ-SPEC VOC Laboratory Sensor Testing Protocol  
    • Presentation by: Xiaobi (Michelle) Kuang, Air Quality Specialist, South Coast AQMD, United States

      The potential for VOC sensors to improve our understanding of VOC exposure in community and fence-line applications is an alluring prospect. While VOC sensors can provide greater time and spatial resolution compared to conventional VOC monitoring techniques, there exists substantial uncertainties in the accuracy of these sensors due to potential interferences from temperature, humidity, and cross sensitivities with other gaseous pollutants. To better understand the operation and function of VOC sensors, as well as their selectivity and sensitivity to various VOC compounds, the South Coast AQMD Air Quality Sensor Performance Evaluation Center (AQ-SPEC) has developed a laboratory testing protocol to evaluate VOC sensors. Here, we present performance evaluation results from an inaugural set of VOC sensors tested under this protocol. The performance of VOC sensors was evaluated using metrics such as data recovery, intra-sensor variability, sensor detection limit, accuracy, and correlation to reference instruments. ANOVA was also performed to understand sensor sensitivity towards different explanatory variables in a simulated outdoor test. These results will help to understand limitations of VOC sensors, provide insights in their potential applications, and may also aid their future development to improve sensitivity and selectivity. This is a companion presentation to “A New AQ-SPEC Laboratory Testing Protocol for VOC Air Quality Sensors”. 
    • Disposable Wireless Two-stage PM Sensor Based on Magnetoelastic Resonators 
    • Presentation by: Zeyu Li, PhD Candidate, University of Michigan, China

      Exposure to fine airborne particulate matter (PM) is associated with many adverse health effects, including impaired pulmonary function, asthma, and cardiovascular diseases. There is a need for small, low-cost portable sensors that can be used for monitoring personal exposure. Most conventional PM sensors are based on light scattering, with limited accuracy and sensitivity. Resonant micromechanical sensors, which depend on the progressive accumulation of particulate matter leading to predictable reduction in their characteristic frequencies, have provided superior performance, but require periodic replacement. Piezoelectric-on-silicon chips for surface and bulk acoustic wave resonators have been reported to provide excellent mass sensitivity but their relatively expensive fabrication processes and wired electrical connections that make their replacement undesirable. Here, we present a low-cost and accurate solution with easily replaceable wireless sensors. It utilizes a polymeric platform chip (36.0×30.0×13.5 mm3) that houses two wireless magnetoelastic resonators – one each for PM10 and PM2.5 measurements – along with aerodynamically tailored fluidic channels and planar metal coils for interrogation of both resonators. The platform chip is designed for simple replacement of the resonant sensor elements. The sensor provides, respectively for PM2.5 and PM10, sensitivities of 0.13 Hz and 0.12 Hz peak frequency variation per min×µg/m3.
    • Insights in performance of VOC sensors under variable conditions
    • Presentation by: Martine Van Poppel, Senior Scientist, Flemish Institute for Technological Research (VITO), Belgium 

      Sensors have been extensively used for regulated pollutants but data on performance of sensors for emerging and unregulated pollutants is not commonly available. Volatile Organic Compounds (VOC) are of concern for environment and health. In some locations - like industrial environments or harbors - spatial variability can be large. Whereas diffusive samplers have been extensively used to assess ambient concentrations of VOCs, sensors can give real-time information at a high temporal resolution which can be useful to identify sources. However, VOC sensors can have issues similar to other gas sensors (e.g. interference of T, RH and other gases) and specific issues related to difference in sensitivity. In this study we evaluated PID based sensor system and compared it to other technologies. 

      In this presentation we will share the insights from lab and field tests of TVOC sensors. We assessed between-sensor-uncertainty, accuracy, repeatability, impact of temperature, RH and interfering compounds. The results will give insight in potential applications of VOC sensors.

    Air sensor network design and evaluation through an equity lens

    Sensor network design decisions, particularly around sensor siting, are influenced by the objectives of the sensor network. In this session, we will discuss sensor network design when the outcome or focus is on better understanding equity/inclusivity indicators while ensuring effective sensor distribution. We welcome abstracts that discuss the motivation and logistics of sensor siting in this context, approaches for identifying equity/inclusivity indicators that both inform sensor network design and that can be assessed using low-cost sensor data, and results from sensor network deployments.

    Session Chairs:

    Naomi ZimmermanUniversity of British Columbia, Yanju ChenCARB, Dr. Tianjun LuCalifornia State University Dominguez Hills

    Presentations:
    • Low-Cost Sensor Networks to Measure Ambient Air Pollution in Rural South India: Sites, Design, and Procedures for the AAM-LASSI Study
    • Presentation by: Manish Desai, Research Fellow, Department of Policy & Ethics, School of Medicine, University of Pennsylvania, Philadelphia, PA, USA, United States

      Directly measuring ambient air pollution in rural regions of low- and middle-income counties, despite growing awareness of the potential impact to population health, has remained technically and financially challenging. Low-cost sensor (LCS) networks may offer a practical means by which to help close the information gap. Motivated by this concern and opportunity, the AAM-LASSI research study undertook one of the first deployments of a LCS network in rural South India. Notably, the project occurred within a context of changing household energy patterns.


      In this presentation, we describe AAM-LASSI’s (1) site selection process and the project's resulting geographic, environmental, and socioeconomic setting; (2) LCS network design and operation from device specification to community engagement; and (3) data collection, curation, and calibration procedures. We also share lessons learned from our field experience in order to assist the development of similar projects, as well as for purposes such as regulatory monitoring or citizen science, within resource-poor rural areas. 

      Furthermore, this presentation provides key background for complementary presentations by the research team covering preliminary findings from AAM-LASSI.
    • Prioritizing Air Quality Monitoring in Underserved and Overburdened Communities: Using Environmental Justice Maps and Location-Allocation for Equity-Directed Siting of Monitors 
    • Presentation by: Edmund Seto, Professor, University of Washington, United States

      Past studies have documented inequalities with respect to the siting of air quality monitoring. Yet, federal programs such as the Justice40 Initiative, and state programs such as the Healthy Environment for All (HEAL) Act in Washington, aim to address environmental injustice by prioritizing direct benefits to underserved communities that are overburdened by pollution. 


      In Washington State there are currently 83 PM2.5 monitoring sites operated by the Department of Ecology. These monitors were located at sites that meet state and federal requirements, which were established before the advent of cumulative impacts and environmental justice maps, such as the Climate and Economic Justice Screening Tool (CEJST) and the WA Environmental Health Disparities Map (WA EHD Map) – maps that currently guide policy and funding decisions. 

      Department of Ecology’s current siting of PM2.5 monitors performs reasonably well in terms proximity to communities that rank high in cumulative impacts on the WA EHD Map: Communities ranked highest (rank 9-10 out of a maximum of 10) are on average 4.4 km away from their nearest PM2.5 monitoring site, while communities that rank lower at 7-8 and 5-6 are on average 7.2 and 8.4 km away from their nearest PM2.5 monitoring site, respectively. 

      Location-allocation is a geospatial method that can identify optimal sites that minimize distances between service sites and population demand. We demonstrate the use of location-allocation to identify a hypothetical siting of PM2.5 monitors in Washington that is based on prioritizing EHD Map rankings. The hypothetical air quality monitoring network results in shorter average distances between communities that rank 9-10 and their closest monitoring site (3.0 km) than the current real-world monitoring network. Moreover, the spatial distribution of the hypothetical sites still exhibit good coverage across the state. Interestingly, approximately 30 of the 80 sites are located in the 3-county Puget Sound region, spaced closely at 5-15 km distances. Such close monitor-to-monitor distances suggest that low-cost sensors may be a cost-effective way to provide the monitoring density needed for equity-directed air monitoring.
    • Rethinking Quality Metrics for Equitable Low-Cost Urban Sensor Networks
    • Presentation by: Alex Cabral, Microsoft Research

      The successful deployment of low-cost urban air pollution sensor networks is dependent on knowing where to place sensor nodes. Sensor network design strategies often optimize for area coverage, which does not account for the social nature of cities. Although some researchers have explored quality metrics that focus on social equity, these metrics assume a large, binary sensing region, where entire neighborhoods are "sensed" when they have just one sensor. Given the effect of urban form on the sensors, it is likely that the sensing regions do not cover entire neighborhoods and are not circular, thus these metrics overstate their coverage. Furthermore, because many social metrics are based on census data, which are gathered based on where residents live, they optimize for equity based on home addresses, and do not consider mobility of urban residents. This talk highlights new low-cost urban air pollution sensor network quality metrics that use accurate sensing regions, account for urban resident mobility, and utilize theories such as socially fair facility location to explore equity across urban stakeholder groups.

    Sensors as a component of urban air quality management planning

    How to combine data from regulatory instruments and low-cost sensors for developing and implementing air quality management plans; Local vs regional air quality infrastructure and exploring how these two paradigms of ‘sensing’ pollution present different challenges & opportunities.

    Session Chairs:

    Lekan PopoolaUniversity of Cambridge, Jin XuCARB, Don CollinsUniversity of California Riverside

    Presentations:
    • Heavy-Duty Vehicle Enforcement with Roadside Emissions Monitoring Devices for CARB’s Clean Truck Check Program: Current Status and Future Plans
    • Presentation by: Hang Liu, Staff Air Pollution Specialist, CARB

      Starting from 2023, the California Air Resources Board (CARB) began implementation of the Clean Truck Check program to reduce emissions from heavy-duty vehicles, in order to meet SIP commitments, ensure vehicles are functioning properly, and improve air quality for the public especially for those in disadvantaged communities. As part of the program, CARB’s Enforcement Division is deploying the Portable Emissions AcQuisition System (PEAQS), a plume-capture roadside emissions monitoring device (REMD) utilizing CO2, Black Carbon and NOx sensors to screen trucks and identify potentially non-compliance vehicles.


      CARB is utilizing REMDs in both unattended and mobile deployments. Unattended units are strategically placed at selected locations to operate autonomously and continuously without the need to station personnel at each site. REMDs are also deployed on mobile platforms to help the field enforcement team to target roadside inspection efforts in communities. With data gathered through REMDs, CARB is implementing a streamlined enforcement process to notify owners of vehicles via mail that are identified as high emitters about the actions they are required to take to demonstrate compliance. This streamlined process is a critical enforcement tool to detect non-compliant vehicles and follow up on repairs with increased efficiency.

      This presentation will provide details of the current implementation of REMDs for the Clean Truck Check program, including statistics, summaries and findings from unattended and mobile deployments, as well as from the streamlined enforcement process. It will also include an overview of CARB’s future plans to expand the unattended deployment network, continue mobile deployments especially in disadvantaged communities, and combine owner submitted On-Board Diagnostic (OBD) data with REMD data for deeper insights at vehicle, fleet and manufacturer levels.
    • Urban Air Quality Intelligence: A Smart City Approach with IoT-Based Sensors for Doha-Qatar 
    • Presentation by: Imran Shahid, Research Associate Professor, Qatar University, Qatar

      Air quality monitoring represents a critical aspect of urban sustainability. In this study, we adopt an exhaustive stance on air quality management, integrating established analytical protocols with cutting-edge sensor network technology to enable seamless and comprehensive monitoring of indoor and outdoor air quality parameters. This paradigm shift is underpinned by the integration of wireless communication technology and the Internet of Things (IoT), facilitating the direct transmission of data streams to a centralized repository. In this study, we used an Arduino-based IoT sensor platform, engineered to capture and quantify key air quality constituents, including particulate matter (PM2.5), nitrogen dioxide (NO2), carbon monoxide (CO) and volatile organic compounds (VOCs). The dataset is stored on a server with a geospatial database for robust data analysis and visualization. Precision validation, achieved through systematic calibration and refinement techniques, consistently yields variances within a 15% margin. The systematic and extensive calibration procedures verify the sensor system's reliability across diverse environmental conditions and pollutant levels. We integrated the sensors network within conventional air quality monitoring network for better spatial coverage. In summary, this study highlights the vital role of IoT-based air quality networks for sustainable development within the urban policy framework.
    • Multi-Scale Sensor Integration for Comprehensive Air Quality Management 
    • Presentation by: Miguel Escribano, Business Development Director, BETTAIR CITIES & DNOTA, Spain

      Air quality management is enhanced by the integration of diverse monitoring systems across Urban, Regional and Semi-Industrial scenarios. This work highlights three recent projects, delivered by the companies Bettair (the manufacturer of the Bettair static node, a Class 1 Sensor Systems as defined in CEN TS 17660-1:2021) and Dnota (distributor of Reference Methods) in collaboration with governmental Air Quality agencies. The three initiatives exemplify the hybrid approach: 1) A Urban Hybrid AQ Network allowed the integration of data from three Reference Method stations and seven Bettair nodes, enhancing local urban pollution mapping 2) A Regional Hybrid AQ Network utilized several mobile Reference stations and Bettair nodes for dynamic regional ozone monitoring and 3) A Port Authority´s Hybrid network, utilized Meteorological, gas chromatography, and Bettair sensor systems to monitor port emissions, addressing both air quality and regulatory compliance. These examples illustrate the potential of combining high-accuracy instruments with cost-effective sensor systems to improve air quality management, despite challenges in calibration and data validation. In all three cases, state-of-the-art standards were applied to validate QA&QC and calculate the uncertainty of measurements. The synergy of these systems offers a detailed understanding of pollution sources and dynamics, essential for effective policy-making and community engagement.
    • Blending Sensor and Regulatory Monitoring Data for Exposure Reduction and Exceptional Event Demonstrations in Southern California
    • Presentation by: Scott Epstein, South Coast Air Quality Management District 

      The South Coast Air Quality Management District (South Coast AQMD) is the regional governmental agency overseeing air quality for the 17 million people across the greater Los Angeles metropolitan area (South Coast Air Basin) and the Coachella Valley. The region has significant air pollution challenges despite aggressive efforts to reduce emissions through regulatory and incentive-based measures. In 2020, South Coast AQMD staff deployed an air quality index (AQI) map that blends regulatory monitoring data with consumer-grade sensor data and simulations from a chemical transport model as the primary tool to advise members of the public to reduce their personal exposures. The science underlying the blended AQI map, other publicly available products designed to help reduce exposure, and the public reception to these tools will be presented. Blended PM2.5 surfaces have also been used to retrospectively analyze smoke impacts from past wildfire events when developing exceptional event demonstrations. Once approved by EPA, when determining National Ambient Air Quality Standards attainment status, these demonstrations can be used to exclude monitoring data that are influenced by events that are not reasonably controllable or preventable. Examples showing how the blended AQI map and its underlying PM2.5 data can support exceptional event demonstrations will be presented.
    • Enhancing the Adoption of Air Quality Sensor Data: Bridging the Usage Gap from Local to Global scale.
    • Presentation by: Nuria Castell, Senior Scientist, NILU, Norway
       

      Our research emphasizes the potential of sensor data in advancing air quality monitoring, identifies challenges at both technological and social levels, and proposes solutions to foster the uptake of sensor data from local to global scales.

      In the past decade, there has been a significant increase in the utilization of low-cost sensor technology for monitoring air quality, particularly focusing on fine particulate matter (PM2.5). These sensor technologies have been predominantly deployed in urban settings, playing a crucial role in smart city development and in Citizen Observatories and Citizen Science projects. The potential of sensor data to serve as a robust additional source of environmental information has grown, thanks to advancements in sensor technologies, reduced costs, and the broader involvement of citizens in monitoring activities.

      However, a pertinent question remains: how can we promote the utilization of sensor data beyond the confines of local projects where it is initially generated? Presently, the adoption of sensor data on a regional, national, or global scale faces impediments arising from both technological and social constraints. Noteworthy technological challenges include the absence of harmonized metadata, open data quality processing techniques, and a lack of interoperability and common semantics among data repositories, hindering the seamless integration of sensor data into larger, comprehensive datasets. On the social front, a key obstacle is the mistrust surrounding sensor data, particularly when collected by non-experts or citizen scientists.

      Addressing these socio-technological constraints is imperative, as it holds the potential to enhance the capability for conducting research and formulating evidence-based policies through the utilization of sensor data.

      The accuracy of low-cost sensors varies significantly between devices and changes over time due to factors such as humidity interference or degradation from aging. Current methods for assessing data accuracy often rely on the co-location of sensors with reference instrumentation. However, this approach is not consistently applicable to citizen-owned sensors and is impractical for large networks comprising thousands of sensors. Additionally, co-location exercises are time-limited and do not offer insights into time-dependent accuracy fluctuations, such as those resulting from weather conditions or sensor aging.

      To address these challenges, we are developing various techniques for automated remote validation, where data accuracy is determined on-site. This validation relies on sensor redundancy (multiple sensors in proximity), information from nearby reference stations, model outputs, and satellite data. Furthermore, to ensure seamless integration and interoperability among datasets, we propose adopting a common terminology for communicating data quality and processing levels. The adoption of common terminologies (metadata) will facilitate the aggregation of sensor data into larger datasets, enabling sensors to contribute effectively to the in-situ component of existing observation systems.

      An illustrative case highlighting the valuable utility of sensor data is its contribution to enhancing our comprehension of health effects related to air pollution. Recent research indicates that low-cost PM2.5 sensors generally exhibit good accuracy when compared to reference-equivalent methods. Among the various air pollutants, particulate matter is of great importance due to the studies revealing stronger links between fine PM concentration and adverse health effects. 

      Current official monitoring sites do not represent concentrations in varying microenvironments and occupational settings, which may be higher. Furthermore, the global distribution of official monitoring stations is uneven, with sparse coverage in the Global South. Even in the Global North, the density of these stations is low, resulting in data gaps that pose challenges for epidemiological studies. The integration of sensor data proves instrumental in providing useful estimates of personal-level exposure and increasing the spatial coverage of data.

      At the local and regional scale, the successful assimilation of sensor data with air quality models has been demonstrated, particularly in Norway. In the city of Kristiansand, we augmented the existing two official PM monitoring sites with 30 static sensors mounted at residents' homes and 10 mobile sensors mounted on bikes. The outcomes of assimilating sensor data with the model highlighted that sensors can capture hyperlocal variations in pollution that models struggle to depict. The combination of these two information sources enables the creation of a high-resolution air quality map in near-real time.

      Currently, ongoing efforts are dedicated to the assimilation of sensor data on a European scale. This involves utilizing the CAMS regional air quality forecast and satellite data from the Copernicus Space Component. The primary goal of these efforts is to provide high-resolution maps, ranging from 1 km to 100m, which can be utilized for research and policy development.

    • Use of sensors and micro emission inventory for hotspot management – A case study of Local Air Management Plan (LAMP) in Delhi, NCR 

    • Presentation by: Swagata Dey, Program Project Manager, Environmental Defense Fund, India

      Mitigation of hotspots in cities is an integral part of air pollution management. While reference grade CAAQMS are used for regulatory monitoring and reporting, sensors are now being used to supplement the data. This study combines data from sensors, to supplement existing CAAQMS, along with micro emissions inventory and dispersion modelling to inform targeted grid-based mitigation strategy: “Local Air Management Plans (LAMP) for Delhi NCR. 

      Sensors were deployed within a radius of 2-3 kilometres around a CAAQMS in the chosen hotspot. Mobile monitoring was carried out across fixed routes to supplement stationary data. Pollution maps indicated variation of PM 2.5 levels from 40 µg/m3 to 110 µg/m3 within a small area of 5 km2, varying almost 2.5 times. The spatial variation in NO2 values were much higher. Emissions were mapped through micro emission inventory with extensive primary surveys focusing on local sources. Dispersion modelling was carried out using AERMOD to discern the spatial variation of pollutants from the emission inventory. Sensor data was used to validate the model and action plan was developed. Sensors were selected based on a rigorous collocation study next to reference grade station. 

      The micro-action plan, complete with scenario analysis, can help city authorities manage hyperlocal sources at the ward level, thus making the problem more manageable. This research contributes to advancement in the use of sensors for data driven action to leverage authorities’ needs.  It can be used for enforcement against gross polluters and larger air pollution policy design in an urban environment

    • Real-Time Remote Emissions Monitoring of Heavy-Duty Vehicles: An Enforcement Tool to Detect High NOx or Black Carbon Emitting Vehicles

    • Presentation by: Isaac Lino, Air Resources Engineer, California Air Resources Board, United States

      Diesel engines are a significant source of particulate matter (PM) and oxides of nitrogen (NOx) emissions in California, particularly affecting environmental and economically disadvantaged communities. A relatively small number of high-emitting vehicles contribute a disproportionate fraction of both PM and NOx emissions compared to the rest of the fleet.

      The Portable Emission AcQuisition System (PEAQS) is a remote sensing device developed by the California Air Resources Board (CARB), which measures NOx, black carbon, and CO2 from passing heavy duty vehicle traffic. PEAQS plays a key role in identifying high emitting vehicles for either NOx or PM for enforcement purposes, supporting California’s Clean Truck Check program. CARB Enforcement uses mobile as well as stationary deployment platforms and includes automated license plate reader technology to identify the passing heavy-duty vehicles. The mobile systems are deployed throughout California at various times and are used to flag high-emitting vehicles for immediate, roadside inspection. 

      An emphasis will be made on the effectiveness of using these screening tools for roadside vehicle inspections conducted by CARB Enforcement—a strategy that has demonstrated specific advantages for communities focused on environmental justice. Trends from the mobile PEAQS deployments over the last several years will be shown, alongside data from roadside inspections by Enforcement. This analysis will highlight changes in emissions and citation rates at strategic deployment locations across California. Additionally, the presentation will discuss the construction and implementation of the mobile deployment platform as a crucial element of the enforcement strategy.

    • Comparative Analysis of Microscale Air Quality Response to Traffic and Built Environment from Machine Learning Perspective via Low-cost Sensor Network

    • Presentation by: Ya Wang, Hong Kong University of Science and Technology

      With millions of people living and working in close proximity to busy roads, exposure to air pollution is a major health risk for residents in Hong Kong. As the city continues to grow (e.g., increase in traffic volume, new buildings under construction), assessing the microscale air quality response to these changes is essential to the development of planning strategies.

       We monitored air quality for a month using a low-cost sensor network near Tsuen Tsing Interchange, a highly congested traffic area in Hong Kong. We proposed a machine learning-based model with introduced Google real-time traffic status and frontal area index to simulate air quality, which outperformed existing models and had a high capability to reproduce hourly pollutant concentrations with R2 > 0.71. The newly introduced indicator FAI played a key role in the model, improving performance by 8%.

      Our monitoring revealed three clusters of temporal air quality characteristics with similar trends. The model showed that air quality in these clusters responded differently to traffic activities and built environments, but similarly to meteorological factors. Using interpretable ML methods, we found built environment impacts played a more important role in TRAP accumulation than traffic activities. High-density built environments caused larger increases in NO2 and PM2.5 concentrations. Our findings suggest planners should balance urban development and air quality by considering local features.


    Weight of Evidence: Building the case for change by combining sensor data with other evidence bases

    How to combine data from regulatory instruments and low-cost sensors for developing and implementing air quality management plans; Local vs regional air quality infrastructure and exploring how these two paradigms of ‘sensing’ pollution present different challenges & opportunities.

    Session Chairs:

    Ethan McMahonEM Environmental Services, Rebecca GarlandUniversity of Pretoria and ANGA, Albert PrestoCarnegie Mellon University

    Presentations: 
    • Low-cost air quality sensors and environmental justice 
    • Presentation by: Douglas Booker, Co-Founder & CEO, NAQTS / Lancaster University, United Kingdom

      Low-cost sensors are often touted as a key tool to promote environmental justice by 1) increasing the spatiotemporal resolution of data to more accurately show how air pollution is equally —or unequally— distributed, and 2) democratising science by enabling citizens to generate their own air quality data to raise awareness of potential problems in their own neighbourhoods (especially where no data already exists). While low-cost sensors are indeed a powerful tool, they are not a silver bullet, and may in fact perpetuate environmental injustice.

      This presentation outlines the different challenges and opportunities of using low-cost sensors related to diverse claims of environmental justice. These insights will provoke critical thought about the environmental justice consequences of how low-cost sensors are used, and, ultimately, the role of how air quality science understands, creates, and communicates air quality knowledge.
    • Sensors, Monitors and Tactics in Air Quality Advocacy
    • Presentation by: Karen Grzywinksi, ACCAN

      ACCAN was founded to advocate for accountability and better regulation of a coking facility which was not properly operated or maintained; residents were suffering with pollution that created health problems and affected quality of life. Using a camera trained on the facility, ACCAN collaborated with Carnegie-Mellon University CREATE Lab to record the pollution 24/7. A website was established to make this information and monitoring data available to the public. ACCAN members also became certified emissions evaluators; published stories of residents who suffered the effects of the pollution in a booklet, "Living Downwind"; attended official County meetings to testify about the pollution; and held large public meetings for residents and local and federal regulators. Some ACCAN members purchased one share of company stock which allowed them to speak at annual shareholders’ meetings. ACCAN pressured the local regulatory agency to install passive air monitors throughout the community. This monitoring was conducted for a full year before the facility unexpectedly shut down. Regulators were convinced to continue the monitoring for another year to produce a retrospective pollution study and to undertake an additional retrospective study of residents seeking emergency respiratory and cardiovascular medical treatment. Research generated by these studies confirmed the harm from the coke facility pollution.
    • Breathing Equity: Unmasking Socio-Economic Disparities in Recreational Air Quality with Low-Cost Sensor Insights
    • Presentation by: Amirhosein Mousavi, Research Associate, University of Southern California, United States

      In the dynamic landscape of Southern California, our study investigates environmental justice concerns associated with recreational air pollution exposure during physical activities, such as running, exercise, and outdoor school physical education classes, focusing on parks and school recreational areas. Utilizing low-cost sensors and validated government stations, the research discerns a marked contrast in air pollution exposure between low and high socio-economic census tracts. Individuals in economically disadvantaged areas experience significantly elevated levels of air pollution during physical activities, exacerbating existing environmental injustices. Importantly, our findings challenge prevailing urban planning practices, suggesting that parks may not be located in the least polluted areas within a community. This raises questions about the evolution of air pollution sources around recreational spaces. The study's implications underscore the urgency for targeted interventions and policy measures to address identified disparities, emphasizing the need for a reevaluation of urban planning strategies. Recognizing socio-economic dimensions becomes crucial for immediate and long-term health outcomes, linking to early-stage dementia, Alzheimer's, and respiratory issues. The imperative lies in ensuring equitable access to clean recreational spaces and mitigating evolving air pollution sources, especially in critical areas of community engagement.
    • High Spatiotemporal Resolution Estimates of PM2.5 in West Africa Using Well-Calibrated Air sensors, Satellite Data, and Machine Learning
    • Presentation by: Daniel Westervelt, Associate Research Professor, Columbia University, United States

      Exposure to ambient fine particulate matter (PM2.5) is a leading environmental risk factor for premature death. In Africa, surface PM2.5 data is sparse, hindering pollution mitigation plans and human health improvement. NASA satellite observations provide near-complete spatial coverage, but their columnar nature is imperfect representations of surface pollution. To estimate PM2.5 concentrations at high spatiotemporal resolution (1 km2, daily) across West Africa over the past two decades, we trained, tested, and fine-tuned a machine learning (XGBoost) model with the following data: PM2.5 from reference-grade and calibrated low-cost monitors, aerosol optical depth from MODIS MAIAC satellite retrievals, five meteorological features from ERA5, and seven tropospheric trace gas column or aerosol property features from TROPOMI/OMI satellite retrievals. Preliminary results show that the model performs reasonably well (r2 = 0.73, mean absolute error = 11 µg m-3) in predicting daily PM2.5 compared to observations in six cities across West Africa. Likewise, we will develop a machine learning model focused on Ghana for more localized epidemiological and environmental justice studies. These novel PM2.5 datasets will enable us to identify and explain long-term trends, cycles (seasonal, weekly, and diurnal), and significant sources of PM2.5 in both urban and rural areas. Data from KHRC Ghana long term birth cohorots and epidemiological datasets in rural Ghana will be leveraged to build African-based exposure-response functions. As air quality monitoring networks expand across Africa, the spatial predictions are expected to improve.
    • Invited Presentation by: Ana Hoffman
    • Presentation title and description forthcoming.

    Next Generation Air Quality Monitoring Networks

    The current paradigm for using low-cost sensors typically involves calibrating and deploying one or more sensors, collecting data, and analyzing these data in isolation to determine air quality at each sensors’ location. This session will feature presentations about how we might move beyond that, and what the next generation of air quality monitoring networks might look like. We encourage presentations about innovative approaches which combine data from multiple sensors across a network, and potentially also external systems such as satellite remote sensing instruments or atmospheric transport and chemistry model outputs, to best use the strengths of these different approaches to build up our overall understanding of air quality. Potential topics include data assimilation or data fusion with low-cost sensor data, air quality forecasting with low-cost sensor networks and meteorological data, and strategies for in-situ performance verification or re-calibration of low-cost sensors using network-level data quality metrics. Note that this session is not designed for presentations on new sensor technologies (these should be submitted to the “future sensor development needs” session) nor for presentations about the use of sensors together with socioeconomic data (these should be submitted to the “weight of evidence” session).

    Session Chairs:

    Carl MalingsNASA, Haofei YuUniversity of Central Florida, Pallavi PantHEI

    Presentations: 
    • Observations of volcanic plume chemistry and exposure in Hawaii using a distributed low-cost sensor network
    • Presentation by: Ben Crawford, Assistant Professor, University of Colorado Denver, United States

      Volcanic eruptions pose a potential air quality hazard for hundreds of millions of people worldwide from de-gassing, tephra, and secondary formation of particulate matter (PM).
      This presentation highlights ongoing observations from a low-cost sensor network (18 stations) deployed on the Big Island of Hawai`i from 2021-2023. During active eruptive phases, ~1,000-3,000 tons per day of sulfur dioxide (SO2) gas is continuously released from the Kīlauea summit. As the plume is transported downwind towards populated areas, the SO2 gas photochemically transforms into fine PM, with potential negative health impacts for Island residents. 

      The low-cost sensor network measures SO2 gas concentrations with electrochemical sensors and PM size distributions using a combination of optical particle counters and nephelometers. Network sensors are calibrated in the field against government regulatory-grade monitors. Long-term, high spatial resolution data from the sensor network allows characterization of the plume’s chemical transition in terms of gas-particle conversion rate and particle size distribution during a range of atmospheric conditions. Additionally, the network enables fine-grained population exposure estimates to both SO2 and PM as the plume chemically evolves.

    • A Machine Learning Approach to Forecasting Ground-Level PM2.5 Concentrations at Low-Cost Sensor and Reference Grade Monitor Locations 
    • Presentation by: Daniel King, Geospatial Data Scientist, Sonoma Technology, Inc., United States

      Short-term exposure to particulate matter smaller than 2.5 μm in diameter, referred to as PM2.5, can cause adverse health effects, making short-term spatiotemporal PM2.5 forecasting an important tool in aiding air agencies, public health officials, and the public in day-to-day decision making. Existing, widely available forecasting data sets are generally derived from course-grid physics-based models, such as National Oceanic and Atmospheric Administration’s (NOAA’s) National Air Quality Forecast Capability (NAQFC). While such models provide critical information about future air quality, the coarse resolution limits the ability to reflect local-scale conditions and the ability to incorporate recently observed air pollution is also limited. By taking advantage of the increasing quantities of observational data available, including from low-cost sensors, and technological advances, machine learning models can augment and enhance the atmospheric physicochemical models used to predict PM2.5 concentrations, capturing non-linear statistical associations in the data to improve upon the physicochemical model forecasts. 
      In this study, machine learning models were used to forecast hourly PM2.5 concentrations up to 48 hours into the future. A multi-year archive of hourly PM2.5 concentrations recorded by AirNow monitors and PurpleAir sensors across the contiguous United States was used to train the model. In addition, multiple datasets were compiled to provide spatial predictor variables for the machine learning model, including meteorological, chemical transport, and dispersion model outputs, as well as indicators of land use, land cover, and land form. After data processing and feature engineering, a machine learning model was fit to the data for each forecast horizon, and cross-validation was used to evaluate the models. The machine learning models had a root mean squared error (RMSE) ranging from 5.25 to 10.23 μg/m3, generally increasing with the forecast horizon. The machine learning models consistently outperformed NAQFC, which had an RMSE ranging from 9.83 to 11.06 μg/m3, demonstrating clear improvement upon the physicochemical model’s forecast and the value of this approach for forecasting PM2.5 concentrations. The trained models can be used operationally to produce fine-scale PM2.5 forecasts across the contiguous United States.
    • CAMS-Net: Sustaining a global international network of networks for obtaining useful, actionable data from air sensors 
    • Presentation by: Daniel Westervelt, Associate Research Professor, Columbia University, Lamont Doherty Earth Observatory, United States

      The Clean Air Monitoring and Solutions Network (CAMS-Net) is a National Science Foundation-funded project aimed at creating an international “network of networks” that will facilitate the exchange of knowledge, ideas, and data in order to improve the usage and application of air sensors. It unites scientists, decision-makers, and local stakeholders in developing new methods and best practices for continuously monitoring, sharing, analyzing, and applying air sensor data in both the Global South and North. This talk will highlight some of the successes, challenges, and lessons learned in sustaining an international network of networks for air sensors. In particular, we emphasize successful capacity building efforts, co-created funded research projects, scholar exchanges, and scientific workshops that have had lasting impact on the air sensor community. The network-of-networks approach we take here is critical to make substantive progress in the Global South. Because air pollution is a complex problem, a multidisciplinary approach involving experts from a diverse set of disciplines is required to adequately address the problem. I will highlight in particular a research effort seeded by CAMS-Net to provide a global co-location dataset that is open access and allowing for more powerful calibration models that are more widely applicable to be confidently developed. 
    • DEVELOPMENT OF AN EARLY WARNING SYSTEM FOR  DETECTION AND LOCALISATION OF WILDFIRES Based ON AIR QUALITY SENSOR DATA 
    • Presentation by: Haris Sefo, Head of Research and Sensor Systems Development, Breeze Technologies UG & HCU University Hamburg, Germany 

      Wildfires pose an escalating global threat due to climate change, with rising temperatures and increased droughts fueling their intensity and expansion into previously unaffected regions. Extinguishing large-scale wildfires is perilous and prolonged, endangering lives and incurring massive economic and ecological costs. These fires release CO2 and disrupt carbon sequestration, compounding the climate crisis. Given the current trajectory of greenhouse gas emissions, a further intensification of wildfires is likely.

      To address this crisis, wildfire prevention must incorporate climate adaptation strategies. Innovative early warning systems are crucial to detect and locate fires rapidly in remote areas and prevent their escalation. The Hamburg-based startup, Breeze Technologies, has developed an air quality sensor-based early warning system that employs AI to distinguish wildfires from other sources, such as traffic or industry. Their system has undergone successful field and laboratory testing in collaboration with the U.S. Department of Homeland Security. Since 2023, the system has been implemented for the first time in pilot projects. Among others, in the USA, Canada and Germany.

      This approach represents a promising step toward mitigating the devastating impact of wildfires in a rapidly changing world. 

    • Changing Idling Behavior through Dynamic Air Quality and Idle Detection Messaging 
    • Presentation by: Kerry Kelly, Associate Professor, University of Utah, United States

      Concentrated vehicle engine idling can cause microenvironments of poor air quality, and areas with high idling, such as schools or hospitals, are frequented by individuals at increased risk for negative impacts from poor air quality.  Vehicle idling also wastes fuel and leads to unecessary greenhouse gas emissions.  Anti-idling signage and education have been used to influence driver idling behavior.  However, these strategies have had limited success.  In this study, we try a different approach – dynamic feedback about air quality and idling status. A similar approach, dynamic feedback on driver speed, has become commonplace and has effectively reduced speeding and traffic accidents in the vicinity of the speed display.  Our study aims to evaluate how dynamic messaging about idling behavior and air quality affects driver idling choices.  
      Our study entailed developing a system that integrates networked low-cost air quality sensor measurements and idling vehicle detection based on video and audio inputs to provide dynamic feedback to drivers. The dynamic feedback system comprises seven air quality sensor nodes, one video camera, six microphones, a central server, and two outdoor displays to provide feedback to the drivers regarding idling status and local air quality. The low-cost air quality sensor nodes measured carbon dioxide (CO2), carbon monoxide (CO), nitric oxide (NO), nitrogen dioxide (NO2), total volatile organic compounds (TVOC), and particulate matter (PM2.5).  The video camera and the six microphones provide information about vehicle motion and engine noise.  The team developed a machine-learning technique to identify idling vehicles from the audio and video inputs. The server then integrated the air-quality sensor measurements and idling status (from the audio and video inputs) to provide real-time feedback about vehicle idling and air quality in messages sent to the outdoor displays.  

      We deployed our dynamic feedback system at a hospital's drop-off zone, where numerous vehicles frequently idle.  This deployment lasted 15 days. Five days showed a static control message welcoming patrons to the hospital on the outdoor displays; five days showed a machine learning idle status-based dynamic message; and five days showed an air quality metric-based dynamic message.  The machine learning algorithm performed well over the 15-day deployment, with an average idling label precision of 78.3% and an average non-idling precision of 96.6% compared to onsite notes.  The dynamic message deployment days resulted in an average decrease in PM2.5 concentration, CO2 concentration, and NO concentration of 0.338 ug/m3, 9.1 ppm, and 2.6 ppb, respectively, when vehicles were present compared to the static control message days. The reduction in combustion emissions suggests that our dynamic messaging system can positively influence driver idling behavior and improve air quality in an area that typically experiences concentrated vehicle idling.
    • How far can low-cost methane sensors take us to?
    • Presentation by: Jiayu Li, Universty of Miami

      Given the critical role of methane emissions in climate change and the limited availability of high-resolution, ground-level data, there is a concerted effort among scientists and engineers to develop affordable sensor networks and monitoring techniques. There are ongoing questions about the efficacy of these sensors and their suitability for ambient measurements. In response, we have developed a cost-effective methane sensing node, which includes two metal oxide (MOx) sensors, alongside humidity and temperature sensors, data storage, and telemetry capabilities. This prototype was deployed next to a standard methane analyzer at two different locations - one outdoor and one indoor - for extensive data collection over several months under varying environmental conditions and methane levels. We examined calibration models to assess our system's performance, particularly its ability to monitor background levels and detect enhancements. The sensors showed moderate accuracy in the 2 to 10 ppm range. However, we observed fluctuations in sensor response over time, potentially due to the presence of other gases. Our comprehensive analysis sheds light on the limitations of these methane sensors and opens up a discussion about the optimal conditions and locations for their deployment.
    • Vertical profiles of particulate matter concentration using low-cost sensor on-board drone over Delhi, India
    • Presentation by: Ajit Ahlawat, Research Scientist, Leibniz institute for tropospheric research, Germany

      Delhi is considered as one of the most polluted cities around the world. Previous studies have reported use of ground-based low-cost sensor (LCS) network over Delhi. However, there is still a significant knowledge gap regarding vertical information of air pollutants over Delhi. Therefore, an airborne campaign using LCS on-board drone was conducted at Indian Institute of Technology (IIT), Delhi to provide vertical distribution of particulate matter (PM) from March 12-23, 2021. 

      The drone carried a payload comprising of PM-LCS along with meteorological sensors for PM mass concentrations, temperature and humidity measurements. An inlet was designed using 3-D printer and a dehumidification unit was used in order to minimize the effect of humidity on particle mass concentration. The LCS performance comparison revealed higher correlation (R2= 0.94) when compared against BAM at ground. In addition, multiple sensitivity tests (i.e. for sensor positioning, propeller effects, impact of humidity with and without dehumidification unit etc.) were performed to evaluate the LCS performance while placed on-board drone. 

      The results indicate that incorporating LCS into innovative platforms is feasible, potentially offering detailed data in vertical dimension. Using PM and gas sensors (VOCs, NOx, O3, CO, etc.) on-board drone could enhance the comprehensive understanding of vertical air pollutants distribution.
    • Combining satellite, BAM, mobile sensors, and purple air monitoring to understand air quality status and trends in Antananarivo, the central highlands, and the southeastern coast of Madagascar.
    • Presentation by: Lovanomenjanahary Marline, Post doctoral fellow, Association Vahatra, Madagascar

      Madagascar, a low-income country, has air quality data problem with little to no monitoring across the country. We measured PM2.5 in Antananarivo, other major cities in the highlands, and the southeastern coastal plains of Madagascar. We used four sources of data: a Beta-attenuated Monitor at the US Embassy in Antananarivo, NASA satellite data calibrated to the BAM, purple air sensors and a mobile 2BTech PAM. We found unhealthy to hazardous levels of PM2.5 in most city centers and along highland area roadways, largely attributable to tail pipe emissions of individual motor vehicles. We found moderate background levels of PM2.5 in rural areas, attributable to charcoal production, brick manufacturing, vehicles and forest fires that accumulate in a general haze over the highlands during the dry season. Coastal towns had good to excellent air quality. In Antananarivo, wet season pollution meets WHO’s 24-hr. standards on most days from January through July, and exceeds the 24-hr. health standard on 75% of days in October and 25% of days in Sept and Nov, especially between 5-8 AM and PM. Because outdoor air pollution contributes to approximately 15% of deaths annually, work is needed to raise public awareness, conduct broader monitoring, and reduce emissions.

    Sensing and Exposure Assessment of Hazardous Air Pollutants

    The advent and widespread use of low-cost sensors have dramatically improved the ability of communities to measure their exposure to criteria air pollutants such as particulate matter, ozone, and nitrogen oxides. However, many underserved communities also face exposure to hazardous air pollutants (HAPs), such as toxic metals and volatile organic compounds (VOCs). This session aims to bring together researchers, community members, industrial partners, and government agencies to discuss the latest advancements and challenges in ambient monitoring of and exposure assessment to HAPs. The session will cover a wide range of topics including, but not limited to: 1) Technological advancements that enable near real-time monitoring of toxic metals and VOCs; 2) Identification of HAPs trends and source apportionment; 3) Use of remote sensing technologies to achieve monitoring of HAPs at high temporal- and spatial resolution; 4) Use of ambient HAPs measurements for exposure assessments and air quality model evaluation.

    Session Chairs:

    Hanyang Li, San Diego State UniversitySerena ChungUS EPA

    Presentations:
    • Exploring Ways to Communicate Mobile Air Toxics Data to Communities
    • Presentation by: Ashley Collier-Oxandale, ATOPS Data & QA Unit Supervisor, Colorado Department of Public Health & Environment APCD, United States

      The Colorado Department of Public Health and Environment employs innovative and advanced monitoring techniques to better understand air toxics trends and impacts across Colorado. Among these monitoring approaches are several mobile platforms, one of which is the Colorado Air Toxics (CAT) mobile monitoring van. The CAT employs chemical ionization time-of-flight mass spectrometry (CI-ToF-MS) and cavity ringdown spectroscopy (CRDS) instrumentation to monitor a wide range of VOCs, including benzene, as well as H2S, and HCN in near real-time while driving. This platform was developed in response to an increased demand for understanding and communicating air toxics exposures within disproportionately impacted communities.. However, there are significant challenges in utilizing this high time resolution data to communicate exposure health risks to the public.

      While the instrumentation on the CAT differs from air quality sensors, both in terms of cost and the specific information provided, there are shared challenges when it comes to communicating air toxics data. From a public perspective, being informed of health risk exposures is a top priority. This presentation will provide an overview of ambient mobile monitoring performed with the CAT, and explore a pilot project that employs several data analysis techniques and visualizations. This pilot project was designed to increase the value of mobile monitoring datasets by utilizing high time resolution data to inform communities about potential exposure risks. Furthermore we will discuss how the lessons of this pilot project might apply to air toxics data from sensors and other lower cost instrumentation. 

    • Multi-pollutant and multi-sensor strategies for understanding the sources and spatial variation of VOCs in an urban oil drilling setting 
    • Presentation by: Caroline Frischmon, Graduate Student, University of Colorado Boulder, United States

      The Las Cienegas oilfield in South Los Angeles sits among the most environmentally burdened communities in California. In addition to HAPs exposure from oil extraction, residents are also exposed to traffic-related HAPs from the major freeway cutting through the community. To better understand how these emissions sources influence the spatial variation of HAPs, specifically VOCs, in the community, we set up a dense network of multi-pollutant air sensors.
      The sensor packages, called HAQ-Pods, include an array of VOC metal oxide sensors. While no individual metal oxide sensor in the array is a perfect fit for quantifying tVOC levels alone, we can combine the unique selectivity of each sensor in machine learning models to better capture overall VOC levels. The HAQ-Pods also contain sensors capable of quantifying CO, NOx, and methane. We use these tracer species to investigate how tVOC levels might be attributed to either oil and gas extraction or traffic-related emissions.

      Using this multi-sensor and multi-pollutant technique, we quantify tVOC and methane levels near, midrange, and further from an active oil extraction site in the Las Cienegas oilfield. These levels are compared to tVOC and methane levels near an idle well, a deconstructed well, and a near-roadway site away from any oil extraction. We use these comparison sites to investigate how VOCs can be apportioned within the community using only lower-cost HAQ-Pod data.
    • Fully Automated and Real-Time Posting of Volatile Organic Compounds Monitoring Results in the Colorado Front Range 
    • Presentation by: Detlev Helmig, Principal, Boulder AIR, United States

      Over the past 15 years, the Northern Colorado Front Range has experienced a remarkable growth of oil and natural gas drilling and extraction, with more than 40,000 wells having been drilled in the area east of the Rocky Mountains and north of the City of Denver. The increasing encroachment of oil and natural gas production into suburban and urban neighborhoods has raised citizens’ concerns about hazardous air pollution from the oil and gas industry. Several regional governments have responded to these concerns by implementing air monitoring of primary and secondary oil and natural gas emissions. Volatile organic compounds (VOCs) have been monitored at up to seven regional monitoring stations in parallel by fully automated gas chromatography (GC) systems at hourly time resolution, generating on the order of 50,000 speciated VOCs sample analyses per year (1). This vast amount of data acquisition and reporting has been accomplished by fully automating sample collection and data processing, including the VOCs preconcentration, GC separation, peak identification, peak integration, and computation of analyte ambient air mole fraction results. Data processing routines incorporate blank subtraction and consideration of monthly adjusted FID carbon response factors. Quantification results are reported as time series graphs and in tabulated format to websites within minutes after the end of the GC run to provide hazardous VOCs pollutant data to the public in real time. This presentation will detail the instrumental components, protocols for the data processing, and experiences made from the real-time posting of the monitoring data to public website portals.
      (1) https://www.bouldair.com
    • The AirPen: A Wearable Monitor for Characterizing Exposures to Particulate Matter and Volatile Organic Compounds 
    • Presentation by: Emilio Molina Rueda, Research Assistant, Colorado State University, United States

      Exposure to air pollution is a leading risk factor for disease and premature death, but technologies to assess personal exposure to air pollutants, including the timing, source(s), and location of such exposures, are limited. We developed a small (200 g), quiet, wearable monitor, called the AirPen, to quantify personal exposures to particulate matter and volatile organic compounds (VOCs) as a function of time, location, source, and activity. The AirPen combines physical sample collection (PM onto a filter and VOCs onto a sorbent tube) with a suite of low-cost sensors (for PM, VOCs, noise, temperature, humidity, light intensity, location, and motion). Dozens of minerals, VOCs, and metals can be quantified through sample analysis. We validated the AirPen against conventional personal sampling equipment in the laboratory and then conducted a series of field demonstrations to measure at-work and away-from-work exposures to PM and VOCs among employees at agricultural and manufacturing facilities. Time-integrated PM2.5 and benzene concentrations collected using AirPens were strongly correlated with (Pearson’s r > 0.94), but slightly lower than, concentrations derived from conventional samplers. Multiple data streams from the AirPen were used to apportion personal exposures to VOCs and PM into specific locations and activities. The different field studies showcase how the AirPen can be used for exploratory assessments of a wide spectrum of pollutants and environments, and for targeted assessments to evaluate compliance with regulatory exposure levels to specific hazards. Furthermore, our results illustrate how the AirPen can be used to advance our understanding of the air exposome by characterizing the context of exposure, even in the absence of detailed activity diary data.
    • A New AQ-SPEC Laboratory Testing Protocol for VOC Air Quality Sensors 
    • Presentation by: Wilton Mui, AQ-SPEC Program Supervisor, South Coast Air Quality Management District, United States

      Sensors for measuring VOCs are of increasing interest for consumers and underserved communities to gauge exposure to non-criteria air pollutants. However, there is relatively limited public information on the performance of commercially-available VOC sensors to guide users toward appropriate offerings on the market. The South Coast AQMD Air Quality Sensor Performance Evaluation Center (AQ-SPEC) has developed a first-of-its-kind VOC sensor laboratory testing protocol. This test protocol evaluates VOC sensors for their ability to measure a wide range of VOC concentrations, their susceptibility to environmental and gaseous interferents, and their behavior in simulated outdoor conditions. Performance of VOC sensors from this protocol is characterized by metrics such as data recovery, intra-sensor variability, accuracy, and correlation to reference instruments. An overview and technical specifics of this new VOC sensor laboratory testing protocol is provided. This is a companion presentation for “Performance Evaluation Results for the Inaugural Set of VOC Air Quality Sensors Tested Under the AQ-SPEC VOC Laboratory Sensor Testing Protocol”. 
    • Laboratory evaluation of a low-cost electrochemical formaldehyde gas sensor 
    • Presentation by: Zheyuan Pei, Graduate Research Assistant, University of Utah, United States

      Formaldehyde is a known human carcinogen and an important air pollutant. However, current strategies for formaldehyde measurement, such as spectrometric and optical techniques, are expensive and labor intensive. Low-cost gas sensors have been emerging to provide effective measurements of air pollutants. In this study, we evaluated eight low-cost electrochemical formaldehyde sensors (SFA30, Sensirion®) in the laboratory with a broadband cavity-enhanced absorption spectroscopy as the reference instrument. As a group, the sensors exhibited good linearity of response (R2 > 0.95), low limit of detection (11.3±2.07 ppb), good accuracy (3.96±0.33 ppb RMSE and 6.2±0.3 % NRMSE), acceptable repeatability (3.46% averaged coefficient of variation), reasonably fast response (131-439 s) and moderate inter-sensor variability (0.551 intraclass correlation coefficient) over the formaldehyde concentration range of 0-76 ppb. We also systematically investigated the effects of temperature and relative humidity on sensor response, and the results showed that formaldehyde concentration was the most important contributor to sensor response, followed by temperature, and relative humidity. The results suggest the feasibility of using this low-cost electrochemical sensor to measure formaldehyde at relevant concentration ranges in indoor and outdoor environments.
    • Measuring the breathable heavy metals in Sacramento City’s air 
    • Presentation by: Wayne Linklater, Professor and Chair, Dept. of Environmental Studies, California State University - Sacramento, United States 

      Aerosolized heavy metals pose a serious risk to peoples’ health. Until recently, however, measurements of exposure to breathable heavy metals has been limited. The technology was too expensive and large for repeated sampling at fine spatial and temporal scales. The recently developed TARTA (Toxic-metal Aerosol Real Time Analyzer) device uses atomic emission spectroscopy to measure metal elements that accumulate on electrodes as air is passed over them. We used the TARTA device to conduct mobile air-quality sampling of stratified-randomly selected sites along 38 routes across Environmental Justice Zones, areas of underserved neighborhoods, in Sacramento County. Sampling occurred between 5am and 10pm on all days of the week over a 12-month period beginning March 2023. No site was resampled to maximize the measurement of air quality spatial and temporal variance. The subsequent dataset provides insights as to the most common and most variable aerosolized heavy metals in a medium-sized urban center, answering some questions but raising new ones about the risk posed by breathable heavy metals.
    • Advancements in Near-source Emission Assessment Data Analysis
    • Presentation by: Megan MacDonald, EPA

      The EPA Office of Research and Development’s Next Generation Emission Measurement (NGEM) program focuses on researching the deployment of lower-cost air sensors near known emissions sources. The air quality measurements collected with these sensors can be used to indicate the presence of potential air pollution due to fugitive emissions and industrial process malfunctions. They can also potentially help reduce impacts to nearby communities. The sensor pod (SPod) is one such lower-cost fenceline sensor technology. It collects time-synchronized meteorological and volatile organic compound (VOC) concentration data at a rate of 1 Hz. It can trigger canister grab samples under elevated VOC conditions, which can then be analyzed to determine the exact components of elevated signal spikes.
      Sensor deployments produce vast chemical concentration and meteorological data requiring Quality Assurance (QA) treatments, analysis, and processing prior to being finalized as a functional data set. Data analysis applications are helpful for automating the processing of large datasets like these. One such application, The SEnsor NeTwork INtelligent Emissions Locator (SENTINEL), was developed to meet this need by providing users with a tool to process, analyze, and visualize data from multiple sensor deployments. This presentation will overview the methods used to develop the QA procedures, data processing, and analysis built into the SENTINEL application. Several case studies utilizing the current and potential features of the application will be shown, including the impacts of different background correction levels, the potential for emissions estimation using back trajectory modeling, and the future integration of other types of sensors into the application interface.


      Disclaimer: This abstract has been subjected to review by EPA ORD and approved for publication. Approval does not signify that the contents reflect the views of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use.

    Clean Air Monitoring and Solutions Network: getting useful, actionable data out of low cost sensors for air quality action

    The Clean Air Monitoring and Solutions Network (CAMS-Net) establishes an international network of networks that unites scientists, decision-makers, city administrators, citizen groups, the private sector, and other local stakeholders in co-developing new methods and best practices for real-time air quality data collection, data sharing, and solutions for air quality improvements. CAMS-Net brings together a vast network of multidisciplinary member networks from all around the globe. The project establishes a mechanism for international collaboration, builds technical capacity, shares knowledge, and trains the next generation of air quality practitioners and advocates, including domestic and international graduate students and postdoctoral researchers. We welcome presentations on topics including but not limited to:

    • calibration/correction factor development 
    • data sharing, management, and standardization
    • application of sensor data to modeling and satellite data
    • role of sensors in source attribution
    • public engagement and capacity building
    • decolonization of science and responsible science
    • public policy and health
    Session Chairs:

    Dan Westervelt, Columbia UniversityAlbert PrestoCarnegie Mellon University

    Presentations:
    • Impact of Seven Years of Open Air Quality Community Data in Pakistan
    • Presentation by: Abid Omar, Founder, Pakistan Air Quality Initiative, Founder, Pakistan

      Cities in Pakistan consistently report extreme levels of air pollution, with Lahore ranking as the most polluted city in the world. At the same time, these cities remain a ‘data gap’ in terms of availability of air quality data with sporadic monitoring by government and by other institutions. Moreover, there is little scientific research interest for Pakistani cities, especially in comparison with other regional LMIC cities that have relatively lower air pollution levels and affected populations.

      This paper is the first comprehensive survey of available data for Pakistan, by publishing seven years of low-cost sensor data collected by a LCS community network for the 4 largest cities in Pakistan, namely Karachi, Lahore, Islamabad, and Peshawar. Statistical comparisons with available reference-standard data and remote sensing data is also done. As this is community-driven data, novel data visualizations are utilized in the statistical analysis to make the results understandable by the general public, which helps in the dissemination of air quality data.

      Challenges in data collection are also covered when working in LMIC city areas affected by access issues, intermittent electricity supply and data outages, as well as the learnings from seven years of community work, and use of sensors for community awareness and education in Pakistan. The impact from this community monitoring network has been tremendous in kick-starting awareness in one of the most air-polluted regions of the world.

    • Exploring the robustness of lower capital cost air sensors for understanding the impacts of location-specific agricultural practices on local air quality in Ghana. 
    • Presentation by: Collins Gameli Hodoli, Postdoctoral Research Associate, University of Georgia, Athens, United States

      This study explores the utility of appropriately calibrated lower capital cost (LCC) air quality monitors in understanding PM2.5 pollution in a range of agricultural settings in Ghana and more widely whether this class of sensor can appropriately be used to support targeted local air quality campaigns. Four LCC nodes (Clarity Node-S) were initially calibrated at the University of Ghana, Accra, using a multiple linear regression model following a 4-week colocation with Teledyne API T640 PM Mass Monitor for PM2.5. The nodes were then deployed for a 16-week period (August 28 to December 19, 2022) at a range of rural agricultural sites around the Fumesua and Sokwai area. Fumesua and Sokwai Farms (FF and SF) were respectively selected as the control and main site with no burning and burning activity. Hourly, we observed highest PM2.5 pollution in the Sokwai Community (SC, ~200 μg/m3), followed by SF (~150 μg/m3), Fumesua Community (FC, ~100 μg/m3) and FF (50 μg/m3). We also observed a significant drop in PM2.5 pollution from October to December at the SF. Comparing the observations to GEOS-CF and MERRA-2 modelled PM2.5, we observed that the models overestimated PM2.5 but GEOS-CF seems to reflect the diurnal patterns slightly better than MERRA-2. The results demonstrate the utility LCC monitors for establishing baseline of PM2.5 pollution in these settings.
    • Using Colocations from 25+ Cities to Create a Global Correction for Optical Low-Cost Sensors 
    • Presentation by: Garima Raheja, PhD Candidate, Columbia University, United States

      Air pollution is a leading cause of global premature mortality. Traditional methods of measuring air pollution are expensive, technically challenging, and inaccessible for many low-income marginalized communities. Advancements in low-cost sensors (LCS) are helping bridge the data gap left by these reference-grade monitors. Novel data science techniques are being used to develop correction factors for LCS, but these studies generally 1. use co-locations with expensive reference-grade monitors 2. utilize temperature, humidity and other measurements to account for variation in hygroscopicity and optical properties and 3. are often local in scope, limited to one city or metro area.

      Can we use correction factors developed in one community, in another? We use colocations contributed by 25 community projects and regulatory studies (including in NYC, Ohio, Accra, Lomé, Kinshasa, London, and Kolkata) at varying climatologies to assess the performance of 4 machine learning techniques, and compare them to correction factors in the literature. Additionally, we develop a Global Gaussian Mixture Regression (GMR) machine learning model trained on co-locations from communities in the Clean Air Monitoring and Solutions Network (CAMS-Net). GMR has proven successful for correcting LCS data: in Kinshasa, the GMR-corrected Purple Air data resulted in R2 = 0.88 when compared to the MetOne BAM1020, and in Accra, the GMR lowered Mean Absolute Error of Clarity data from 7.51 𝜇g/m3 to 1.93 𝜇g/m3.

      We find that in most cases, the Global Gaussian Mixture Regression model performs 57-96% as well as using a local correction model, which means that using this model could provide high levels of accuracy in a community using LCS without the need for a $100,000+ reference monitor colocation.In some cases, such as Nairobi, the Global Gaussian Mixture Regression model is actually 1.2x better than using correction developed with local colocation.

      Global GMR is greater than the sum of its parts: contribution from some communities has reciprocated progress in many more. We present an open-source dashboard that enables the correction of data from 20,000+ PurpleAir and Clarity sensors around the world without a reference monitor colocation, and has allowed community groups, regulators and policymakers around the world to make the most of their LCS data.
    • Evaluation of Spatial and Method-Related Sources of Bias and Error in Low-Cost PM2.5 Sensors with Continuously Varying PM2.5 Composition 
    • Presentation by: Daniel King, Sonoma Technologies, United States

      Low-cost sensors can contribute to a range of air quality monitoring objectives, including measuring and attributing local-scale air pollution trends, filling in gaps in the regulatory monitoring network, and providing very-low latency information on air quality current conditions. However, debate continues among regulatory agencies, scientists, and the public regarding the suitability of measurements from these instruments for these purposes. Instrument performance may degrade under specific conditions, with dust and smoke events noted as cases where for which high levels of bias have been observed that widely used sensor correction approaches may not properly correct. Such findings generally rely on comparing sensors to nearby monitors of known quality. These methods provide vital insights into measurement errors, but are limited as they do not quantify the relative contribution of sensor-related and spatial errors (due to time-coincident differences in ambient concentrations or conditions at “nearby” monitors). There is a need to understand and provide context to these relative contributions and identify potential opportunities to advance the appropriate use of low-cost sensors. 

      In this work, we developed a new statistical assessment of sensor uncertainty that facilitates the quantification of spatial and sensor errors under varying PM2.5 composition or size fraction conditions. We used a sliding window analysis to quantify the relative contributions of spatial and sensor errors to overall error across distance bins. This approach allows us to quantify the absolute error and bias due to sensor characteristics alone. Through the use of PM2.5 composition data, we further quantified the relationship between PM2.5 source type and the sensor performance. With these relationships, we considered existing and potential new sensor calibration models for the conterminous United States that adjust raw sensor data based on PM2.5 event type. We benchmarked the calibrations using a large-scale historical dataset. The results of our work provide valuable information on PM2.5 data applications and appropriate uses in real time to adjust for varying conditions, with explicit quantification of estimated error and bias. These results can support data fusion applications that provide real-time and forecasted air quality information to decision makers and the public.


    Air Aware Schools: Protecting Students and Enhancing Learning with Outdoor Air Quality Monitoring

    This panel discussion will shed light on the importance of monitoring outdoor air quality within educational institutions. Join us as our panelists explore the transformative role of real-time air quality data in safeguarding the well-being of students and staff while simultaneously enhancing the overall learning environment.

    As schools face an increasing frequency of wildfires alongside persisting air pollution challenges, the integration of air quality sensors across school districts has proven invaluable for informed decision-making during emergency events. The accessibility of these data to parents and students further demonstrates the potential for outdoor air quality monitoring to promote healthy learning environments and enhance the curriculum for educational institutions.

    Session Chairs:

    Ryan Higgins, Clarity & Sarah Kroening, Children's Health Alliance of Wisconsin


    Advancing Wildfire Monitoring with Air Quality Sensors — How Air Sensor Data Are Being Integrated at the Local and National Scale

    In this panel discussion, we delve into the use of air quality sensors for wildfire monitoring, with a focus on enhancing public safety and response strategies during wildfire events. Our panel of experts will share their invaluable insights on the deployment of low-cost air quality sensors — at both the local and the national level — to address the distinct challenges posed by wildfires.

    Air quality sensors that offer durability, ease of deployment, and real-time data can be invaluable for the development of effective wildfire smoke monitoring and alerting systems. While acknowledging the challenges associated with some low-cost sensors, our discussion will highlight the potential for reliable sensor technology in wildfire monitoring, especially in scenarios where government agencies can lead the way in implementing necessary changes.

    Session Chairs:

    Ryan Higgins, Clarity


    Looking beyond London: How Breathe Cities Will Leverage Low-Cost Sensors for Air Quality Management Globally

    Join us for a discussion on the Breathe London model's adaptability and its implications for reshaping the future of air quality management in cities around the world.

    The Breathe London air quality monitoring program — led by the expert team at Imperial College London’s Environmental Research Group — is a success story that offers a flexible and applicable framework for achieving cleaner air in urban areas globally.

    In this panel discussion, we'll share lessons learned from deploying more than 450 air quality sensors (and growing) across London — including how to build the awareness and partnerships that breathe life into the program. We’ll examine the critical elements that have led to the program’s sustained success — including the operational mode of the team running the network, funding, technical support, data sharing, community engagement, and capacity building — and share how the Breathe London model offers a blueprint for cities looking to establish robust, high-resolution air quality monitoring programs worldwide.

    Session Chairs:

    Ryan Higgins, Clarity, Iq Mead, Imperial College London


    Air Sensors in South Asia — Advances and Updates

    There has been tremendous growth in the use of low(er) cost air quality sensors in low- and middle-income countries, especially in South Asia. Presentations during this session will describe sensor deployments and their intent and highlight both successes and challenges and share data from deployments to date. We expect abstracts focusing on both technical details of deployments and on applied findings from these deployments. This session will both share findings from existing deployments and share learnings with those seeking to deploy these types of networks around the world.

    Session Chairs:

    Ajay Pillarisetti, UC Berkeley, R. Subramanian, CSTEP, Swagata Dey, EDF