<p>Aerosol concentrations over Europe and Germany were simulated for the years 1985 and 2013 using the aerosol-chemistry transport model COSMO-MUSCAT. The aerosol fields from the two simulations were used in a high-resolution meteorological model for a sensitivity study on cloud properties. The modelled aerosol and cloud variables were compared to a variety of available observations, including satellites, remote sensing and in-situ observations. Finally, the radiative forcing of the aerosol could be estimated from the different sensitivity simulations.</p><p>Due to reduction of emissions the ambient aerosol mass and number in Europe was strongly decreased since the 1980s. Hence, today&#8217;s number of particles in the CCN size range is smaller. The HD(CP)<sup>2</sup> (High Definition Clouds and Precipitation for Climate Prediction) project amongst others aimed at analysing the effect of the emission reduction on cloud properties.</p><p>As a pre-requiste, the aerosol mass, number, and composition over Germany were simulated for 1985 and 2013 using the regional chemistry-transport-model COSMO-MUSCAT. The EDGAR emission inventory was used for both years.</p><p>The model results were compared to observations from the two HD(CP)<sup>2</sup> campaigns that took place in 2013 (HOPE, HOPE-Melpitz) as well as the AVHRR aerosol optical thickness product, which is available from 1981 onwards. Despite the fact, that emissions of the 1980s are very uncertain, the modelled AOD is in good agreement with observations. The modelled mean CCN number concentration in 1985 is a factor of 2-4 higher than in 2013.</p><p>Within HD(CP)<sup>2</sup>, the ICON weather forecast model was applied in a configuration allowing for large-eddy simulations. In these simulations, the time-varying CCN fields for the year 1985 and 2013 calculated with COSMO-MUSCAT were used as input for ICON-LEM. In the present-day simulation, the cloud droplet number agrees with observations, whereas the perturbed (1985) simulation does not with droplet numbers about twice as high as in 2013. Also, for other cloud variables systematic changes between the two scenarios were observed.</p>