Christian Asker

A recently developed air quality modelling system is used to produce high-resolution concentrations (with street-canyon effects not included) of air pollutants (NO2, PM2.5,PM10 and O3) over all of Sweden. As a next step the population exposure is quantified and associated health impacts and socioeconomic costs for the years 2023 and 2030 are calculated.

The concentrations of NO2 are predicted to decrease from 2023 to 2030 due to lower traffic exhaust emissions. Modelled exceedances of the air quality standard for the yearly mean of NO2 are limited to the nearest vicinity of a few large roads in 2023, while there are no exceedances for 2030 (excluding street-canyon effects). The average population exposure of NO2 decreases from 4.0 μg/m3 to 2.9 μg/m3.

PM10 and PM2.5 concentrations in 2030 are predicted to remain on a similar level as in 2023, with a possible slight upward trend. The reasons for increasing concentrations are more traffic and thus higher levels of road, tyre and brake wear particles, as well as slightly higher regional background levels. Modelled exceedances of of the air quality standard for the yearly mean of PM2.5 are not expected in neither 2023 nor 2030, while PM10 exceedances are limited to the vicinity of larger roads in both years (again excluding street-canyon effects). The average population exposure of particles shows verylittle variation between 2023 and 2030 (from 8.1 to 8.2 μg/m3 for PM10 and 4.6 to 4.7 μg/m3 for PM2.5).

O3 concentrations are predicted to stay at similar levels in 2030 compared to 2023, with reductions in rural areas and locally increased concentrations close to larger roads. Thethe average population exposure changes only slightly from 57 μg/m3 to 56 μg/m3

The following health impact assessment is based on NO2, local and regional PM2.5, respectively, and wear particles in the PM10 fraction, combining the high-resolution exposure modelling with concentration–response functions from recent epidemiological evidence.

For 2023, the total number of incident mortality cases is estimated at approximately 5 966 (95% CI: 3 905–8 035), with contributions from regional background PM2.5, locally generated PM2.5, wear particles (PM₁₀), and NO2. Compared with previous assessments, the methodology distinguishes more clearly between these sources, including a separate treatment of wear particles using PM₁₀-based risk estimates.

The analysis also includes a broader set of morbidity outcomes. Large numbers of cases are estimated for functional outcomes, including ~680 000 (579 000–782 000) work loss days and ~48 000 (15 700–80 400) school absence days, showing substantial impacts on productivity. Among chronic diseases, estimates include 1 083 COPD cases, 720 dementia cases, 374 diabetes cases, and 251 stroke cases annually. Additional outcomes newly included in the assessment comprise autism spectrum disorder (ASD) (1 422 cases) andallergic rhinitis (148 cases), based on emerging evidence for air pollution effects in children.

In the scenario work for 2030, in total a lower estimated mortality attributed to air pollution was estimated (5 627 cases, e.g. 6% lower), however both of the cases attributed to local wear particles and regional background PM2.5 were higher (621 vs 554 and 3601 vs 3467 cases, respectively).

Methodologically, mortality estimation has been refined by separating PM2.5 into regional background and locally generated components, while wear particles are assessedusing PM₁₀ exposure–response functions. This adjustment addresses the contribution of coarse particles from non-exhaust traffic emissions and avoids underestimation of their health effects, although the uncertainty due to limitations in source-specific risk estimates remains.