Spatio-temporal modelling of individual exposure to air pollution and its uncertainty

We developed a generic spatio-temporal model to quantify individual exposure to air pollution, using personal activity profiles derived from GPS and diaries, ambient air quality, and an indoor model. To enhance accessibility and reusability, the model approach is deployed as a web service. The model is applied to estimate personal exposure towards PM10 and PM2.5 for ten individuals in Münster, Germany. Modelled daily averages range for PM10 between 17 and 126 and between 6 and 84 μg m−3 for PM2.5. Comparison with personal monitoring data shows good agreement at temporal resolutions from 5 min to one day. Uncertainties in the model results are considerable and increase with higher exposure levels. Large deviations between modelled and measured exposure can often be explained by missing data on indoor emissions or insufficiently detailed activity diaries. The developed model allows the assessment of individual exposure with uncertainties on a high spatio-temporal resolution. By providing the methodology through a web service interface and using generic indoor parameter distributions, the model can be easily transferred to new application areas or could be provided for public use to identify hazardous exposure events.

► We developed a generic individual spatio-temporal exposure model. ► The model shows good compliance with validation measurements. ► The model approximated exposure better than urban background station measurements. ► We provide the model functionality through an open web service interface.