Using Commercially Available Low‐Cost Monitors to Estimate the Hourly Spatial Variability of Particulate Matter Concentrations across a Metropolitan Area

In U.S., the National Ambient Air Quality Standards (NAAQS) set the limit values for six principal “criteria” air pollutants including PM2.5. Data are primarily collected to assess the citywide air pollution concentrations for regulatory purposes. PM2.5 is measured at one or a few urban stations within major cities or in rural locations. This sparse spatial resolution is insufficient to capture the intra-urban spatial variability of air pollution that is driven by the locations and strengths of local sources, the effect of street canyons and complex terrain, and urban heat island effects. Consequently, exposure misclassifications are likely to occur when using these data for epidemiological studies. In addition, NAAQS for PM2.5 requires the attainment of annual or daily limit values. However, recent studies have reported associations between high hourly PM2.5 peaks and mortality/morbidity, particularly cardiovascular events [1]. Consequently, it is important to increase the temporal resolution to capture air pollution peaks responsible of short-term health outcomes. The accessibility of low-cost sensing for air pollution may be a valuable resource to improve the spatial and temporal resolution of current routine monitoring networks. Low-cost monitors (LCMs) are much less expensive than scientific-grade instruments, physically smaller and lighter (generally portable), collect data with high time resolutions (from few seconds to minutes), require less maintenance, and have low power demands. However, they are not designed to meet rigid performance standards, and they produce data with much less accuracy than scientific-grade instruments. Thus, LCMs require careful calibration and post-processing of data. Recently, we have used data collected with commercially available LCMs at multiple locations across a metropolitan area of the eastern U.S. (Rochester, NY) during two consecutive winters (2015–2016 and 2016–2017). These monitors (Speck, Airviz Inc., Pittsburgh, PA) were tested under laboratory [2] and field [3] conditions. Data were also used to predict the hourly small-scale variability of PM using sophisticated land-use regression models [4]. The results of a summer/fall sampling campaign (June to October 2017) that essentially completes our dataset to cover all the seasons over three years (2015 to 2017) will be described. Forty-nine LCMs placed in weatherproof cases were deployed outdoors at residential locations in Rochester NY, while another unit was co-located to the NYS DEC air quality monitoring site. Raw data were originally collected at 1 min time resolution. Data were handled to return robust and reliable datasets at 1 h resolution time. Instrumental biases were assessed during 3 days of field co-location with a GRIMM 1.109 aerosol spectrometer pre and post-field deployment. Multiple pairwise analyses were used to investigate the collected data, including coefficient of divergence and signed rank tests of the value distributions. The data were affected by a large but correctable bias that was caused by the low PM concentrations typically measured in Rochester. However, this main limitation was overcome by a careful instrument calibration and validation of data prior to and after the sampling campaigns to ensure unbiased datasets. Despite the lower accuracy of data, results show that the use of these monitors provides the opportunity for successfully improving the spatial resolution of particulate pollution. [1] Gardner, B. et al., 2014. Ambient fine particulate air pollution triggers ST-elevation myocardial infarction, but not non-ST elevation myocardial infarction: a case-crossover study. Particle and Fibre Toxicology 11(1), 1. [2] Zikova, N., et al. 2017. Evaluation of new low-cost particle monitors for PM2.5 concentrations measurements. J. Aerosol Sci. 105, 24–34. [3] Zikova, N., et al., 2017. Estimating hourly concentrations of PM2.5 across a metropolitan area using low-cost particle monitors. Sensors 17, 1922. [4] Masiol, M., et al., submitted. Hourly land use regression models based on low-cost PM monitor data