Long-Term Performance of Low-Cost PM2.5 Sensor Calibration Utilizing Denver's Love My Air Network
TEHYA STOCKMAN, Shelly L. Miller,
University of Colorado Boulder Abstract Number: 75
Working Group: Urban Aerosols
AbstractDenver, Colorado is operating a network of over 30 low-cost PM2.5 sensors (Plantower, Lunar Outpost, and Clarity Movement Co.), with the oldest sensors brought online in 2018. While sensor networks are growing in popularity and many corrections have been developed for low-cost PM2.5 sensors, there is a gap in understanding how these low-cost sensors and corrections perform over multiple years. Most studies developing sensor corrections are less than one year in length. Denver’s network, which includes eight co-located sensors at six federal equivalent monitoring sites since 2018/2019, presents an opportunity to evaluate performance over multiple years. We evaluated sensor bias and variability over time and across sensor manufacturers. We also developed and evaluated corrections for the sensor network using multi-linear regression and machine learning models. We considered models that included variables for meteorology, smoke events, dust events, and sensor age.