ETC/ATNI Report 2020/8: Understanding Air Quality Trends in Europe. Focus on the relative contribution of changes in emission of activity sectors, natural fraction and meteorological variability.

This report examines air quality trends in Europe based on a combination of in-situ observations available in the AQ e-reporting database, air quality chemical transport models (CTM) and statistical modelling/machine learning using Generalized Additive Models (GAM). We focus on providing more insight on how meteorological variability, natural factors, and emission from the main activity sectors might have influence the development in PM10, NO2 and ozone concentrations between 2000 and 2017. Documenting the long term evolution of air quality in Europe allows us also to understand the extent of improvement of detrimental impacts on human health or ecosystems. Emission changes are the main driver of all air pollutant trends. For NO2 and PM10, both the GAM and the CTM results indicate that emission changes contribute to at least 90% of the 2000-2017 trend. For ozone peaks (as 4MDA8), meteorology can be important. The GAM model estimates that it contributes to an increase counteracting mitigation effort up to a magnitude of 20 to 80% (compared to the effect of emission and background changes) in Austria, Belgium, Czech Republic, France, and Italy. Given the good skill of the GAM model to capture meteorological effect, this estimate can be considered quite robust. The relative contribution of agriculture and industry to the total PM10 mass has been reduced by around 30% for both sectors, but the similarity of evolution is not directly linked to the emission trends in the respective sectors. The relationship between emissions and concentrations is nonlinear and depends on availability of precursor gases to form ammonium sulphate and ammonium nitrate. The relative contribution of traffic sources to PM10 has been reduced with around 20%, while the trend attributed to residential heating is marginal. The heating sector has become a relatively more important contributor to the aerosol pollution and needs more attention. The model also indicates that the natural contributions (such as sea salt and dust) has had little impact on the long-term changes in PM10. The analysis includes observational data only from stations with data available for at least 14 years in the period 2000-2017. This drastically reduces the number of monitoring sites included in the analysis and the spatial representativity of the assessment, with bias towards countries benefiting from a long-term monitoring network. Further improvements of models as well as observational basis are needed to reduce the uncertainties. Understanding organic aerosols from the residential heating sector should be a priority.OCP/EEA/ACC/18/001-ETC/ATN