Calibration of PurpleAir PA-I and PA-II Monitors Using Daily Mean PM2.5 Concentrations Measured in California, Washington, and Oregon from 2017 to 2021

Abstract

Large quantities of real-time particle data are becoming available from low-cost particle monitors. However, it is crucial to determine the quality of these measurements. The largest network of monitors in the United States is maintained by the PurpleAir company, which offers two monitors: PA-I and PA-II. PA-I monitors have a single sensor (PMS1003) and PA-II monitors employ two independent PMS5003 sensors. We determine a new calibration factor for the PA-I monitor and revise a previously published calibration algorithm for PA-II monitors (ALT-CF3). From the PurpleAir API site, we downloaded 83 million hourly average PM2.5 values in the PurpleAir database from Washington, Oregon, and California between 1 January 2017 and 8 September 2021. Daily outdoor PM2.5 means from 194 PA-II monitors were compared to daily means from 47 nearby Federal regulatory sites using gravimetric Federal Reference Methods (FRM). We find a revised calibration factor of 3.4 for the PA-II monitors. For the PA-I monitors, we determined a new calibration factor (also 3.4) by comparing 26 outdoor PA-I sites to 117 nearby outdoor PA-II sites. These results show that PurpleAir PM2.5 measurements can agree well with regulatory monitors when an optimum calibration factor is found.

Indoor contribution to PM 2 .5 exposure using all PurpleAir sites in Washington, Oregon, and California

“…19 The hourly outdoor data shown above were averaged to create 3.5 million daily averages. Because our work in our companion paper 16 showed that the PA-I and PA-II sensors had the same calibration factor, we combined the 3190 indoor PA-I indoor sensors with the 981 PA-II indoor sensors and calculated 1.3 million daily indoor averages to compare with the outdoor levels (Figure 2). The median (IQR) infiltration factor was 0.24 (0.13, 0.40)…”

“…Therefore, in a companion study to this one, we determined a new CF for the PMS 1003 sensor, using nearby regulatory monitors in three states (Washington, Oregon, and California). 16 The state of California was chosen because it has the highest number of PurpleAir monitors for any state, accounting for 75% of all PurpleAir monitors. The adjoining states of Oregon and Washington were chosen since all three states have similar climates dominated by the Pacific Ocean and also all three states experienced large forest fires in recent years, providing extensive data on the highest concentrations likely to be experienced.…”

Long-Term Indoor-Outdoor PM2.5 Measurements Using PurpleAir Sensors: An Improved Method of Calculating Indoor-Generated and Outdoor-Infiltrated Contributions to Potential Indoor Exposure

“…The ALT-CF3 method has been found to provide less bias, better precision, and a lower limit of detection than either the CF1 or CF_ATM algorithms [ 47 , 48 ]. This method is presently available on the PurpleAir mapping page at the PurpleAir Website ( , accessed on 1 January 2023) as an alternative to the proprietary algorithms offered by Plantower.…”

“…The following supporting information can be downloaded at: , (References [ 30 , 44 , 45 , 46 , 47 , 48 , 49 , 55 , 56 ] are cited in the Supplementary Materials).…”

Spatial Variation of PM_2.5 Indoors and Outdoors: Results from 261 Regulatory Monitors Compared to 14,000 Low-Cost Monitors in Three Western States over 4.7 Years

“…A previous study collected publicly available hourly PM2.5data from ~160 regulatory sites and particle number concentrations from ~10,000 outdoor and ~4,000 indoor PurpleAir monitors in the states of Washington, Oregon, and California covering the 4.7year period from Jan 1, 2017 to September 8, 2021 1 . For the PurpleAir sites, PM2.5 concentrations were calculated using an improved algorithm based on particle numbers in three size categories as reported by the PurpleAir monitors.This algorithm is completely independent of the two proprietary algorithms provided by Plantower, and has been shown to have reduced bias, improved precision, and a lower Limit of Detection (LOD) [1,8,[26][27][28]. The algorithm is called ALT-CF3 and is available on the PurpleAir main page as one of 5 "conversion factors" that can be chosen instead of the Plantower proprietary algorithms.…”