Particulate matter source identification and triangulation using a low-cost sensor network in Nigeria

Knowledge of air pollution sources is crucial for air pollution mitigation and policy making. Due to multiple constraints, air quality monitoring is often limited in lower- and middle-income countries, where rapid urbanisation and industrialisation are causing the local population to be exposed to dangerous concentrations of pollutants. The use of low-cost sensor (LCS) networks offers a more affordable alternative and a stepping stone in monitoring and assessing the air quality where research-grade equipment is currently unavailable. The present study re-assesses a previously published dataset of particulate matter (PM) concentrations in Ile-Ife, Nigeria, for source identification using k-means clustering and triangulation of the sources. The method identified six clusters based on particle number concentration profiles, separating daytime and night-time conditions, as well as the periods when local or regional sources had the greatest effect on the local air quality. The highest concentrations of particulate matter in Ile-Ife were associated with stagnant conditions and at the sites closer to the industrial activities in the area, for which the PM concentrations were about 10 times higher than the daily average (mean ± standard deviation PM_2.5 30.6 ± 40.7 µg m^-3, and PM₁₀ 83.4 ± 259.2 µg m^-3). The methodology also provides information on the circumstances when cleaner conditions occur, and these were found during the daytime with higher wind speeds from the southwest (SW), when PM concentrations were significantly lower than the average at all sites (PM_2.5 11.5 ± 4.7 µg m^-3, and PM10 37.2 ± 26.0 µg m^-3), but still above the annual limits recommended by the World Health Organization (WHO). Additionally, the polluting effects of the urban background as well as from traffic were identified, and the atmospheric conditions which increased the importance of these sources were elucidated.