PurpleAir sensors overcount PM2.5 in humid smoke, so personal AQI readings are wrong
environmentenvironment0 views
Individuals and families who invested $200-250 in PurpleAir outdoor sensors to get hyperlocal AQI readings during fire season are getting systematically inflated PM2.5 numbers because PurpleAir's Plantower laser particle counters cannot distinguish between wildfire smoke particles and water droplets that condense on smoke particles in humid conditions. The EPA's correction factor (the 'ALT-CF=1' formula on the AQI map) partially adjusts for this, but only on the EPA's own map overlay — not in the PurpleAir app itself or in home automation triggers people set up. A family in Portland might see AQI 280 on their PurpleAir sensor, panic, seal their house, and run a $40/month portable air purifier on max when the true AQI is 160. The structural reason this persists is that PurpleAir uses $15 Plantower PMS5003 sensors that physically cannot do humidity correction at the hardware level, and replacing them with a nephelometer would 10x the device cost.
Evidence
Barkjohn et al. (2021) in Atmospheric Measurement Techniques found PurpleAir sensors overestimate PM2.5 by 30-90% at relative humidity above 60% during wildfire events compared to Federal Reference Method monitors. The EPA developed the ALT-CF=1 correction factor specifically to address this but it is only applied on the EPA AirNow Fire and Smoke Map, not in PurpleAir's own interface or API by default.