Improving ammonia emissions in air quality modelling for France

We have implemented a new module to improve the representation of ammonia emissions from agricultural activities in France with the objective to evaluate the impact of such emissions on the formation of particulate matter modelled with the air quality model CHIMERE. A novel method has been set up for the part of ammonia emissions originating from mineral fertilizer spreading. They are calculated using the one dimensional 1D mechanistic model “VOLT'AIR” which has been coupled with data on agricultural practices, meteorology and soil properties obtained at high spatial resolution (cantonal level). These emissions display high spatiotemporal variations depending on soil pH, rates and dates of fertilization and meteorological variables, especially soil temperature. The emissions from other agricultural sources (animal housing, manure storage and organic manure spreading) are calculated using the national spatialised inventory (INS) recently developed in France. The comparison of the total ammonia emissions estimated with the new approach VOLT'AIR_INS with the standard emissions provided by EMEP (European Monitoring and Evaluation Programme) used currently in the CHIMERE model shows significant differences in the spatiotemporal distributions. The implementation of new ammonia emissions in the CHIMERE model has a limited impact on ammonium nitrate aerosol concentrations which only increase at most by 10% on the average for the considered spring period but this impact can be more significant for specific pollution episodes. The comparison of modelled PM10 (particulate matter with aerodynamic diameter smaller than 10 μm) and ammonium nitrate aerosol with observations shows that the use of the new ammonia emission method slightly improves the spatiotemporal correlation in certain regions and reduces the negative bias on average by 1 μg m−3. The formation of ammonium nitrate aerosol depends not only on ammonia concentrations but also on nitric acid availability, which is often a limiting factor in rural regions in France, and on meteorological conditions. The presented approach of ammonia emission calculation seems suitable for use in chemistry–transport models. ► NH3 emissions from mineral fertilizer spreading in France are simulated using a 1D mechanistic model. ► Emissions display high spatiotemporal variations depending especially on soil pH and rates and dates of fertilization. ► The comparison of modelled PM10 with observations shows that the spatiotemporal co