Improving PurpleAir PM2.5 Sensor Accuracy in Cold Conditions: Developing Tailored Correction Factors
JENNO JOSEPH-LEENOSE-HELEN, Srijan Aggarwal, Raghu Betha, Dominique Pride, Alana Vilagi, University of Alaska - Fairbanks
Abstract Number: 461
Working Group: Instrumentation and Methods
Abstract
Low-cost PM2.5 sensors have gained significance among air quality experts and citizen scientists due to the growing understanding of the harmful impacts of PM2.5 on human health. Renowned for their relatively high accuracy and affordability, PurpleAir (PA) sensors serve as reliable tools for measuring PM2.5 concentrations. Despite the development and application of various correction factors to enhance PA sensor accuracy, none adequately address the challenges posed by extreme cold conditions, particularly in interior Alaska. This study examines the efficacy of existing correction factors when applied to wintertime raw PA data from the city of North Pole, Alaska, revealing their inadequacy in low temperatures. Consequently, the research underscores the necessity for tailored correction factors specific to localized environments, especially at low temperatures. Upon identifying the shortcomings of existing models in low-temperature settings, we develop a new correction factor accounting for performance at low temperatures (0 to -25°C). The raw PA data is compared to PM2.5 concentrations reported by Beta attenuation monitoring (BAM) detectors to develop the correction factor. The study also revealed that PA sensors cease to function below -25°C and prolonged exposure to cold affects their performance, including the accuracy of any correction factors in predicting PM2.5 concentrations. Implementing a monthly correction factor may offer improved accuracy in correcting raw PA data. The proposed correction factor aims to empower residents of cold climates to utilize low-cost sensors for accurate measurement of localized PM2.5 concentrations, enabling informed decisions to mitigate exposure risks.