<p>Air pollution is one of the main causes of damages to human health in Europe with an estimate of about 380 000 premature deaths per year in the EU28, as the result of exposure to fine particulate matter (PM2.5) only. In this work, we focus on one specific region in Europe, the Po basin, a region where chemical regimes are the most complex, showing important non-linear processes, especially those related to interactions between NO<sub>x</sub> and NH<sub>3</sub>. We analyse the sensitivities of PM2.5 to NO<sub>x</sub> and NH<sub>3</sub> emissions by means of a set of EMEP simulations performed with different levels of emission reductions, from 25 % up to a total switch-off of those emissions. Both single and combined precursor reduction scenarios are applied to determine the most efficient emission reduction strategies and quantify the interactions between NO<sub>x</sub> and NH<sub>3</sub> emission reductions. The results confirmed the peculiarity of secondary PM2.5 formation in the Po basin, characterised by contrasting chemical regimes within distances of few (hundreds of) kilometres, as well as strong non-linear responses to emission reductions during wintertime. One of the striking results is the increase of the PM2.5 concentration levels when NO<sub>x</sub> emission reductions are applied in NO<sub>x</sub>-rich areas, such as the surroundings of Bergamo. The increased oxidative capacity of the atmosphere is the cause of the increase of PM2.5 induced by a reduction in NO<sub>x</sub> emission. This process can have contributed to the absence of significant PM2.5 concentration decrease during the COVID-19 lockdowns in many European cities. It is important to account for this process when designing air quality plans, since it could well lead to transitionary increases in PM2.5 at some locations in winter as NO<sub>x</sub> emission reduction measures are gradually implemented. While PM2.5 responses to NO<sub>x</sub> and NH<sub>3</sub> emission reduction show large variations seasonally and spatially, these responses remain close to linear, i.e. proportional to the emission reduction levels, at least up to −50 % because secondary aerosol formation chemical regimes are not modified by those relatively moderate ranges.</p>