Advancing Regional Air Pollution Exposure Assessment: Revealing Underestimated Long-Term Exposure to PM2.5 in Greece

The estimation of human exposure to air pollution presents well-known methodological challenges. Two major challenges are incorporating population activity and accounting for outdoor pollutant concentrations infiltrating indoor environments. These aspects are often overlooked in current exposure assessments at urban and regional scales, introducing biases that result in non-representative exposure estimates and associated health effects.

In this study, we present a method for regional dynamic exposure estimation by integrating population activity and the infiltration of air pollutants into indoor environments. Using the time-microenvironment-activity concept in combination with open-source datasets for the spatial and temporal distribution of the European population, our framework generates grids of population activity across different microenvironments (home, work, schools, transport, etc.) with a spatial resolution of 1x1 km² and a temporal resolution of 1 hour for all of Europe. Each microenvironment is assigned seasonal, literature-based outdoor-to-indoor infiltration factors. The resulting dynamic population grids can be created for any urban or regional to domain in Europe and can be applied to air pollutant concentrations derived from any air quality model that match these domains.

To illustrate the impact of regional dynamic exposure estimates compared to static estimates (based on residential addresses), we combined the dynamic population activity data with a high-resolution (1x1 km²) gridded dataset of PM_2.5 concentrations across Greece from 2015 to 2022. This dataset was produced using a recently developed and extensively evaluated Random Forest modeling approach to downscale regional to urban pollutant concentrations specifically for Greece. Initial results reveal that regional-scale dynamic estimates incorporating population activity resulted in >10% higher mean exposure to PM_2.5 compared to exposure estimates based on static populations but can vary significantly between different microenvironments. These findings suggest that current population exposure estimates to ambient air pollution across Europe are likely to be underestimated.