Long-term outdoor comparison of low-cost particulate matter sensors in an urban environment and potential uses

Bulot, Florentin (2019) Long-term outdoor comparison of low-cost particulate matter sensors in an urban environment and potential uses. Public Health England - 2019 Annual UK Review Meeting on Outdoor and Indoor Air Pollution Research : Current and future perspectives on air quality and public health, Surrey County Cricket Club, London, United Kingdom. 16 May 2019.

Abstract

Introduction: Exposure to particulate matter (PM) is one of the leading causes of morbidity and mortality worldwide. Understanding personal exposure requires a higher spatiotemporal resolution than provided by current reference stations. Low-cost PM sensors may facilitate improved spatiotemporal resolution of PM monitoring, but there are uncertainties regarding the reliability of their data. This work aims to characterise the in-field performance of four models of low-cost PM sensor.
Method: Three PM monitoring boxes, each containing one of each of four models of low-cost PM sensors (2xPlantower, Alphasense, Honeywell), and temperature/relative humidity sensors, were installed at each of two schools in Southampton. Sensor performance was assessed by comparing reported PM concentrations between sensors within the same box, across sites, and with a background reference monitoring station. Effects of PM concentration and meteorological factors on sensor performance were analysed and response to short-lived events was observed.
Results: Sensors present good agreement between each other (Pearson r>0.9, p<0.0001). Inter-model differences were observed including temporal drift for some models of sensors and different behaviour against relative humidity. All models present similar variability with background reference concentrations. Response to spikes of pollution also varies across sensor models. Low-cost PM sensors are suitable to detect and track short-lived events of pollution when used as a network as well as tracking hourly variation of pollution and obtained moderate to good agreement with reference instruments for hourly data (0.61<r<0.88, p<0.0001).
Conclusion: Inter-model variability suggests that sensors should not be deployed individually. Variations between different models of sensors imply that different correction methods should be applied for different models of sensors. The low-cost PM sensors analysed are suitable to track short-lived events of pollution and deployment of a dense network of these sensors may provide useful information for personal exposure assessment and interventions evaluation across urban environments.

This record has no associated files available for download.

Contributors

Download statistics