Calibration details by pollutant

How Clarity calibrates air quality sensors for PM2.5, NO₂, Ozone, and Black Carbon to ensure precise pollutant measurement across diverse conditions.

PM2.5 Calibration

Global PM2.5 Calibration

Clarity’s Global PM2.5 Calibration is designed to enhance sensor performance universally. The latest version (v2) leverages data from over 625 Clarity Monitors and 98 regulatory-grade reference sites in 84 cities. The Global Calibration for PM2.5 achieves a median R² of 0.79 and an RMSE of 2.7 µg/m³, delivering highly reliable data by correcting for aerosol composition variability. This calibration is preset for all Clarity devices, ensuring accuracy even when collocation-based calibration isn’t feasible. More details can be found here: link.

Collocation-Based PM2.5 Calibration

To tailor PM2.5 measurements to local conditions, Clarity offers collocation-based calibration. This process involves deploying Clarity Nodes near reference-grade monitors for at least four weeks, collecting data on pollutant concentrations, and environmental conditions. Using an 80/20 data split for model development and testing, Clarity builds a project-specific model that improves accuracy by accounting for regional particulate characteristics. Cross-validation and quality assurance checks further enhance model robustness. As part of this work Clarity provides sensor-specific accuracy metrics.

NO₂ Calibration

Global NO₂ Calibration:

The Global NO₂ Calibration corrects baseline shifts and interferences caused by temperature and humidity, common with electrochemical cell (ECS) sensors. This calibration leverages a machine learning model (LGBM decision tree regressor) trained on collocation data from various global sites (450 collocated Node-S devices, 2000 collocation months, 45 cities in different climates) to provide a generalized adjustment profile that stabilizes sensor readings and aligns them closely with reference monitors. The result is more consistent and accurate NO₂ data across varying climates, enabling reliable measurements even without site-specific calibration.

Collocation-Based NO₂ Calibration

For high-accuracy NO₂ monitoring tailored to specific projects, Clarity offers a collocation-based calibration approach. This process begins with a four-week collocation period, during which Clarity Nodes collect data alongside reference monitors to capture local pollutant variability. For NO2 it is very important to choose 4 weeks with ambient temperatures and relative humidities that cover a representative range of year-round conditions (this usually means a fall or a spring month). After QA/QC checks, Clarity applies additional data filters, excluding outliers and focusing on typical environmental conditions. An 80/20 data split is used for training and testing, with repeated cross-validation to build a reliable model specific to each sensor. In addition to compensating for sensors baseline shifts as the Global NO2 calibration does, this method reduces device-to-device variations and provides quantifiable performance metrics (R² and RMSE) for each sensor.

Ozone Calibration

Clarity’s Ozone Modules utilize FEM-grade optical detection technology rather than low-cost sensors, ensuring inherently high data accuracy. Due to this advanced detection principle, additional calibration methods are unnecessary; factory calibration provides precise, reliable measurements. For continued accuracy, Clarity offers a factory recalibration every two years as part of its Sensing-as-a-Service package.

Black Carbon Calibration

The Black Carbon Modules also employ sophisticated optical detection technology, allowing for precise measurements without extensive calibration requirements. Factory calibration, performed every two years and included in Clarity’s Sensing-as-a-Service, is sufficient to maintain high data accuracy over time.