Better representing the spatial and temporal variability of ammonia emissions is a major challenge for improving operational tools of air quality prediction (Prev'Air). This issue, supported by public policies, concerns both background pollution and spring peaks of particulate pollution. To this end, the Amp'Air project uses the spatialized dynamic inventories of ammonia emissions generated with the Cadastre_NH3 tool as input for atmospheric chemistry-transport models. The originality of this tool relies on the combination of a process-based model of ammonia volitilization in the field, Volt'Air, with geo-referenced and time-explicit data: land use by crops, soil distribution and properties, meteorological conditions, and, more specifically, the description of nitrogen fertilization by crop type (types of fertilizer, dates, doses, methods of application and abatement...).
