Low-Cost Investigation into Sources of PM2.5in Kinshasa, Democratic Republic of the Congo

Abstract

Despite having a population of 16.3 million, Kinshasa, Democratic Republic of the Congo (DRC), has had little attention toward air quality monitoring. We deployed a MetOne Beta Attenuation Monitor (BAM-1020) for reference PM2.5 and a QuantAQ Modulair, the latter of which includes measurements of gas-phase NO2, O3, CO, and CO2, in addition to PM1, PM2.5, and PM10. Here we present the first results from this aggregated, multisensor, multispecies network in DRC. We first compare the Modulair against the BAM-1020, finding an r2 of 0.76 and a mean absolute error (MAE) of 6.97 μg m–3 (hourly data). We develop a correction factor using multiple linear regression, improving MAE to 5.54 μg m–3. We leverage gaseous pollutant concentrations, particle size distribution data, and anemometer data to draw conclusions about the sources of PM2.5 in Kinshasa. We link factors resolved from a non-negative matrix factorization method using the gaseous and particle bin concentrations to source profiles. We find a 3-factor solution that points to a CO-dominated, supermicron particle source indicative of secondary particles from local combustion, along with a submicron particle-dominated source indicative of primary particles from combustion and a regional biomass burning source. Our results highlight the need for the implementation of clean air solutions in the DRC.