Uncertainties in sector/source contributions. D6.3

The Copernicus Atmosphere Monitoring Service (CAMS) currently offers a number of products to support decision and policy makers regarding mitigation of air pollution at the European scale (https://policy.atmosphere.copernicus.eu/). These policy products provide information about the main sources and drivers of air pollution, through so-called source apportionment or attribution (SA) or source receptor relationships (SR). A key factor influencing the variability in the CAMS source-receptor products is the diversity of source attribution methods. The current service employs three regional chemical transport models (CTMs)—LOTOS-EUROS, EMEP, and CHIMERE— each applying different source attribution techniques: tagging (contribution estimation), brute force (impact of 15% emission reductions), and surrogate modelling (impact
of variable emission reductions). Due to the nonlinear chemistry involved in secondary aerosol formation, these methods can yield different results. This study assesses the consistency and comparability of source-specific PM contributions across the different CTMs and source attribution methods. To this end we have conducted comparative experiments across the three models (EMEP, LOTOS-EUROS, CHIMERE) and different source attribution methodologies (brute force (BF), local fraction (LF), tagging/labelling (TS) and non-linear surrogate modelling (SM)) using a harmonised set-up in terms of inputs (emissions, meteorology, boundaries) and setup (resolution and domain). Additionally, model results were compared with observational data from Positive Matrix Factorisation (PMF) and specific tracers to identify key areas of agreement, divergence, and
opportunities for improvement.