2018-03-08: Field calibration of electrochemical NO2 sensors in a citizen science context

We assessed the performance of low-cost air quality sensors which were used to measure air pollution in Amsterdam during the 2016 Urban AirQ campaign. We show that the current generations of electrochemical NO2 sensors may provide useful complementary data on local air quality, provided that the experiment is properly set up and the data are carefully analysed.
In many urban areas the population is exposed to elevated levels of air pollution. However, real-time air quality is usually only measured at few locations. These measurements provide a general picture of the state of the air, but they are unable to monitor local differences. New low-cost sensor technology is available for several years now, and has the potential to extend official monitoring networks significantly even though the current generation of sensors suffer from various technical issues.

Citizen science experiments based on these sensors must be designed carefully to avoid generation of data which is of poor or even useless quality. This study explores the added value of the 2016 Urban AirQ campaign, which focused on measuring nitrogen dioxide (NO2) in Amsterdam, the Netherlands. Sixteen low-cost air quality sensor devices were built and distributed among volunteers living close to roads with high traffic volume for a 2-month measurement period.

Each electrochemical sensor was calibrated in-field next to an air monitoring station during an 8-day period, resulting in R2 ranging from 0.3 to 0.7. When temperature and relative humidity are included in a multilinear regression approach, the NO2 accuracy is improved significantly, with R2 ranging from 0.6 to 0.9. Recalibration after the campaign is crucial, as all sensors show a significant signal drift in the 2-month measurement period. The measurement series between the calibration periods can be corrected for after the measurement period by taking a weighted average of the calibration coefficients.

Validation against an independent air monitoring station shows good agreement. Using our approach, the standard deviation of a typical sensor device for NO2 measurements was found to be 7 µg m-3, provided that temperatures are below 30 °C. Stronger ozone titration on street sides causes an underestimation of NO2 concentrations, which 75 % of the time is less than 2.3 µg m-3.

Read the full article here.
Comparison of a sensor's NO2 time series with the nearby Oude Schans station (8-day snapshot), and the effect of bias correction. For comparison, measurements of Vondelpark station (were the sensor was calibrated) are also shown.
Distribution of residuals of NO2 measurements between the sensor device and Oude Schans station during the campaign period, with and without bias correction.