Practical Guidance for Using PurpleAir Particle Monitors for Indoor and Outdoor Measurements in Community Field Studies

Abstract Low-cost particle monitors have been widely evaluated in laboratory and ambient monitoring settings, but we have less knowledge about their performance for long-term indoor and outdoor monitoring in residential settings. We seek to provide practical guidance for using a type of low-cost particle monitors that have become widespread for indoor and outdoor monitoring in community field studies, PurpleAir PA-II monitors. We base our insights on experiences in a community-led residential field study in and around homes of predominantly agricultural workers in California’s San Joaquin Valley. Our guidance spans three categories: (1) providing tools for handling and merging disparate data structures resulting from Wi-Fi-transmitted data and data collected on onboard microSD cards, (2) assessing performance metrics of PA-II monitors from laboratory co-location and field measurements, and (3) assessing data collection success rates of Wi-Fi data transmission and microSD card data acquisition from our study locations. The post-processing methods we demonstrate can successfully align data from both Wi-Fi transmission and microSD cards. Laboratory co-location measurements demonstrated that > 90% of the tested monitors performed well relative to each other (high precision), with only a few problematic monitors that warranted further investigation or exclusion from use. The application of co-location factors generated using the mean of all co-located monitors as a reference did not significantly affect distributions of field-measured indoor or outdoor PM2.5 concentrations. Relying solely on Wi-Fi data transmission in our study would have resulted in large data loss (i.e., < 50% success rate); using microSD card storage with PA-II-SD monitors increased the data collection success rate to over 80% in these settings. This work contributes to the growing body of knowledge on low-cost particle sensor performance and usability. Graphical abstract