Ten different combinations of ammonia and nitrogen oxides emission reduction scenarios have been investigated. This analysis has been performed through the application of the chemical transport model (CTM) FARM on two monthly periods characterised by significant measured levels of ammonia and particulate matter, respectively during Spring and Autumn 2011. The simulation domain covers the whole Po Valley (hereafter P-V) and the Alpine Region, with a horizontal resolution of 4 km and 16 terrain-following vertical levels, irregularly spaced from the ground to 10000 m above surface level. The CTM was driven by meteorological input fields produced by the prognostic non-hydrostatic meteorological model RAMS; initial and boundary conditions (IC/BCs) have been derived from FARM national scale simulation at 12 km horizontal resolution. The base case emissions have been prepared merging high resolution bottom-up emission inventories developed by all the P-V administrative Regions applying the same methodology. Local inventories have been combined with the Italian national inventory and EMEP emission data for the portion of surrounding countries inside the simulation domain. Results obtained by the base case simulation (i.e. considering the actual emissions) have been compared with observed levels of PM 2.5 , PM 10 , ammonia, nitrogen oxides and ozone at different monitoring stations located in P-V. The amount and combination of emissions reduction needed to effectively reduce secondary PM levels has been estimated from the analysis of the ten different scenarios performed for the two investigated periods. The role of nitrogen oxides and ammonia on particle mass and composition is confirmed, claiming for the implementation of proper emission control strategies at regional level. INTRODUCTION P-V is one of the air pollution hot spots of major concern in Europe where the EC air quality standards are presently not attained. It is characterised by peculiar topographic features, being surrounded on three sides by the Alps/Apennines chain, and by one of the most densely populated area in Europe, with a global population of about 20 million clustered in different urban areas. Due to its anthropization level and to the presence of intense agricultural and breeding activities, this area is characterised by very high emissions and concentrations of ammonia, whose role as aerosol precursor together with NO X and SO 2 is well known. Since PM composition analysis showed that a large fraction of aerosol has secondary origin, and SO 2 emissions have significantly decreased during the last decades, leading to year-round low concentrations recordings, it is of major interest to understand the potential PM concentration reduction reachable through NO X and ammonia emission limitation policies. According to the Lombardia regional emission inventory (www.inemar.eu), 96% of NH 3 year emissions in the region comes from agriculture activities while the remaining portion (about 4%) is given, particularly in urban areas, by road traffic, from vehicles equipped with three-way catalytic converters. ATMOSPHERIC MODELLING SYSTEM The atmospheric modelling system (AMS) used to simulate the chemical and physical processes involving the pollutants in the atmosphere is based on Flexible Air quality Regional Model (FARM, Silibello et al., 2012). This model has been applied with the SAPRC-99 gas-phase chemical mechanism (Carter, 2000) and the AERO3 modal aerosol scheme implemented in CMAQ framework (Binkowski and Roselle, 2003). The AMS includes modules to reconstruct atmospheric flows and related turbulence parameters and to apportion data from the emission inventories to grid cells.