Field calibration and performance evaluation of AirSensEUR sensor systems in a European pilot project

This report presents the field calibration results of low-cost sensor systems including several pollutant and axillary sensors, AirSensEUR (ASE), at a rural site and three European cities. An automatic calibration procedure has been developed using co-located data of ASEs and reference measurements of air pollutants after QA/QC and outlier filtering procedures. The major findings about the calibration models and their performances in prediction are listed below: 1. The CO-A4 sensors calibrated with simple linear models predicted concentrations at any site and season with high likelihood of meeting the data quality objectives (DQO) of the European Air Quality Directive at the limit values (LV) for the indicative measurements. 2. The NO-B4 sensors calibrated using exponential models with air temperature met the DQO the when the air temperature was close to the ones during the calibration. 3. The calibrated NO2-B43F sensors met the DQO at the LV for the prediction periods with air temperature being similar to the calibration time. 4. The OX-A431 sensors calibrated to measure O3 met the DQO at the LV for a limited time after the calibration. These sensors drifted over time about 6 months, which may require seasonal calibrations. 5. The PMS5003 sensors calibrated with Köhler models generally meet the DQO at the LV with an increasing performance from PM10 to PM1. 6. The OPC-N3 sensors calibrated with Köhler models usually did not meet the DQO except for PM2.5 (and PM1) for the limited periods after they were calibrated.