Estimating background concentrations of PM2.5 for urban air quality modelling in a data poor environment

Atmospheric dispersion models are widely applied to simulate pollutant concentrations such as PM2.5 for use in long- and short-term health studies. A significant proportion of PM2.5 originates outside urban areas in which many people live. It is important to reflect this ‘background’ component in the modelling process in order to provide an accurate representation of the total pollution load experienced by human populations. To be credible, model outputs must be verified against available monitoring data, which, in the case of PM2.5, may be limited to a small number of monitoring sites across a large urban area. Here we evaluate four different approaches to representing background PM2.5 in an atmospheric dispersion model (ADMS-Urban) for Nottingham, UK. A directional approach, based on multiple urban background monitoring sites located outside the study area provides the most robust estimates. Our adopted approach allows us to model both short- and long-term air quality conditions, whilst accounting for local- and regional-scale variations in the pollution burden, and will ultimately enable us to assess short- and long-term effects of air pollution on health. •Background data in local scale air quality models represent regional PM2.5.•Accounting for wind direction in models improves background estimates.•Proximate sources close to monitoring sites influence model verification.•Improved model outputs will aid assessment of acute and long-term health impacts.