2350 - Field Testing a Low-Cost Sensor for Long-Term Measurement of Traffic-Related PM2.5

Abstract

Background Health departments and urban planners have growing needs for high resolution data on urban air pollution, to quantify existing health burdens at the neighborhood scale, to identify and prioritize exposure reduction strategies for pollution hot-spots, to track progress in achieving air quality-related health improvement goals, and to assess health co-benefits of longer-term carbon mitigation strategies. To-date, however, few tools have been available to inform these high priority urban health objectives. As part of the NASA Health and Air Quality Applied Sciences Team (HAQAST) project, this study was designed to field test a unique low-cost sensor for measuring long-term concentrations of PM2.5 mass and composition at sites impacted by road traffic in three cities: Boston, New York, and San Francisco. Methods Starting in August 2017, we deployed University of North Carolina (UNC) passive samplers at 2–3 sites each in Boston, NYC and San Francisco. Sites were chosen due to availability of co-located sampling for PM2.5 using conventional, high-cost sensors, and proximity to major roadways. Each sampler had a filter placed under a fine mesh screen of 1.5 cm in diameter; particles settled gravitationally and by diffusion. We tested two sampling periods: 4 weeks and 12 weeks. Particles accumulated on the filter for as long as the sampler was left open to the air. Samplers were sent to a laboratory for analysis by electron microscopy and analyzed data was returned to us. We also obtained publically available data from high cost regulatory monitors for the sites in Boston, and compared daily averages. Results We deployed samplers on schedule and found that strong winds, particularly near highways, would occasionally disrupt the horizontal orientation of the sampler or cause the sampler to blow away; this was solved by using more sturdy mounting brackets. Initial laboratory results indicate that the 12-week sampling period provides optimal filter loadings. Sampling in 2016 and 2017 at two sites in Boston tell us that the near-road site presents higher PM2.5 annual concentrations by 0.5 to 1 ug/m3. Quantitative analyses comparing the low-cost sensor data with co-located standard devices will be carried out in the coming months as data accumulate. Conclusions Preliminary qualitative results suggest that the novel low-cost UNC sensor can be easily deployed at the community scale, with the potential to achieve broader engagement of the civil society and develop tools to enable our stakeholders to carry out health-based air quality assessments at relevant community scales.