Evaluating collocation performance of low-cost sensors in a semi-urban air quality environment

This study assesses the performance of low-cost air quality sensors (LCS) through collocation with reference-grade monitors in a semi-urban setting. The objective was to evaluate the accuracy and variability of LCS over different collocation durations—1, 3, 7, and 12 days—by comparing their measurements with those from Federal Reference Method (FRM) and Federal Equivalent Method (FEM) monitors.

Using a methodology adapted from the U.S. Environmental Protection Agency (USEPA), the research examined sensor outputs for particulate matter (PM10, PM2.5, PM1), gases (SOx, NOx, CO, O₃), and environmental parameters (temperature, humidity, pressure). The results showed that:

• Correlation (R²) improved with longer collocation periods, stabilizing around 7–12 days.

• Mean Absolute Error (MAE) was highest for 1-day data and reduced significantly in the 7–12-day range.

• Inclusion of meteorological parameters (temperature and humidity) enhanced the performance of LCS data.

The findings suggest that a 7–12-day collocation period is optimal for calibration and correction of LCS in similar environmental contexts.