Quantification of population exposure and health impacts associated with air pollution

Sammanfattning: There is substantial evidence that air pollution, in particular particulate matter (PM), affects our health. The overall objective of this thesis is to understand and quantify population exposure to ambient air pollution and related health impacts. In four included papers, atmospheric dispersion modelling is used to estimate source-specific contributions to pollution levels. Results from the dispersion modelling are evaluated by comparison with available concentration measurements and used to estimate the related health impact. New approaches for health impact assessment are proposed and new methods for exposure assessment are developed. Assessment of health impacts related to different sources of air pollution can be used to identify the most cost-effective abatement strategies. The proposed methods for health impact assessment are applied to evaluate three viable abatement strategies for Stockholm and Gothenburg, the two largest cities in Sweden.The first two papers in the thesis more specifically investigate the importance of using source-specific associations between exposure and premature mortality. One of the main conclusions is that different health risk functions should be applied for near-source and long-range exposure to fine PM. The praxis of today, to use the same linear risk function regardless of source and composition of the PM, most likely underestimates the importance of local sources and may thereby discourage cities from acting to reduce emissions. It is also concluded that a more specific risk function for exposure to coarse PM from road wear would allow for better prioritizations between different abatement measures, especially for countries where studded winter tires are used.Near-source exposure in the urban environment is characterized by strong gradients, requiring a relatively high spatial resolution to capture variations within the population. A new method that allow estimating source-specific exposure with sufficient spatial resolution over large areas and long time periods is presented and applied to create a uniquely detailed NOx exposure assessment for Sweden over three decades.Emissions from road traffic and residential wood combustion (RWC) are found to be the most important contributors to near-source exposure to PM. When three abatement strategies for road traffic are evaluated for Stockholm and Gothenburg, it is shown that a strategy resulting in overall traffic reduction, such as introduction of congestion charges, is a good choice. Reduced use of studded tires is also evaluated, but this measure mainly affects the emissions of coarse PM from road wear, for which health impacts are more uncertain. The on-going electrification of light vehicles also has more uncertain health benefits, at least when studded tires are used, given that the heavier electric vehicles lead to increased emissions of coarse PM from road wear.Improving the description of exposure to PM from RWC is identified as first priority to increase the accuracy in estimates of near-source exposure in Sweden. A comparative study of how this exposure is estimated in the Nordic countries is presented. It is concluded that a more extensive reference dataset with descriptions of emissions and concentration measurements for RWC would be highly valuable for model evaluation and further improvement of model parametrizations.