New calibration preserves air-sensor fluctuations better
One peer-reviewed study examined a new calibration method for inexpensive nitrogen dioxide sensors across four field campaigns at the same Antwerp reference station. Aligning the distributions of measurements preserved the scale of pollution fluctuations better than matching individual timestamps, while conventional calibration retained an edge on some pointwise errors. Corrections were fitted and assessed on the same datasets. Independent testing at other stations is needed before wider claims about reliability.
Science··Morning
Four measurement periods in Antwerp
Four field campaigns from 2022 to 2025 at one Antwerp reference station compared calibration methods for inexpensive nitrogen dioxide sensors. The researchers analysed hourly averages from two sensor units. Temperature, humidity, other gases and ageing can change a sensor’s response outdoors, so a good laboratory fit need not retain the same accuracy in the field. The peer-reviewed study tested a method that aligns distributions of measurements instead of pairing individual readings at the same timestamp.[1]
The scale of fluctuations was preserved
The new method matched the means, spread and selected percentiles of sensor and reference readings. It preserved the scale of pollution fluctuations better than conventional timestamp-paired calibration. Conventional calibration retained an edge on some pointwise error measures. Capturing the size of an excursion differs from locating its exact time. Distribution alignment still needs reference readings from a comparable period; it does not replace the reference station.[1]
Independent validation is needed
Each campaign required a different correction; the study did not establish one universal setting. Some distribution descriptors became more stable after roughly a month of hourly readings, yet the researchers did not determine when recalibration should recur. They built and assessed corrections with the same measurement sets. A limited sensor set at one station cannot establish performance in other cities, seasons or longer deployments. The study measured neither health outcomes nor pollution sources across the city.[1]
Related columns
For more information on this topic, you can read the related columns.
