Nephelometer-based air sensor networks systematically underestimate PM2.5 in dust-influenced West Africa

Abstract Particulate matter air pollution is a major public health burden in West Africa, a region impacted by long-range transport of desert dust in addition to local combustion sources. While new data networks are emerging, most of the measurement growth has come in the form of optical air sensors, more specifically those that behave as imperfect nephelometers. We present an ongoing multi-year measurement campaign since 2024 of 16 monitoring sites across four major metropolitan regions in Ghana spanning the south coast, middle forested region, and the northern savannah, using the QuantAQ Modulair-PM. Results indicate a north-south geographic disparity, where coastal areas such as Sekondi-Takoradi, Accra, and Tema have lower concentrations and infrequent exceedance of EPA Ghana national ambient air quality standards (15% of days or less) while central and northern cities including Kumasi and Tamale have higher concentrations and frequent exceedance (30–49% of days). We find a strong seasonality in ambient PM 2.5 with dry Harmattan season PM 2.5 concentration double or triple wet season levels depending on the site and metropolitan area. Leveraging the Modulair-PM co-located nephelometer (Plantower PMS5003), optical particle counter (Alphasense OPC-N3), and locally calibrated product (which shows low bias and high correlation compared to Federal Equivalent Methods), we assess the bias in Plantower-only estimates of PM 2.5 against the calibrated estimates. We find that more than 30% of the actual PM 2.5 concentration is missed during dusty conditions in Plantower-only sensor approaches, which are the most commonly used in the current literature, while for PM 10 91% of the mass is not accurately observed. This calls into question previous estimates of PM 2.5 and PM 10 in the region or potentially in other dusty environments that rely only on nephelometry, highlights the need for sensors that can accurately size particles, and emphasizes the importance of careful performance evaluation and calibration in deployment of air sensor networks.