Simulations of convective phenomena with the limited area version of the ICON prediction model and comparison with weather radar observations.

The aim of this thesis is to evaluate the performance of the regional version of the numerical weather prediction model ICON in simulating severe con vective storms. Observations from weather radars, complemented by radio soundings and ground weather stations, provide the target for the performance evaluation. The motivation for this analysis stems from the difficulties encountered by high-resolution weather forecast models in detecting convective cells, especially those producing severe thun derstorms, which prevents the timely issue of weather warnings to the public. The supercell thunderstorm that occurred over central Piedmont in the afternoon hours of 6 July 2023 has been selected as case study. Four re-forecasts of this event were considered, each one differing by the combination of microphysical parameterisation scheme and assimilation technique used. The first run mimics the operational settings implementing a bulk-moment microphysics scheme and no assimilation; the second one adopts a double-moment microphysics scheme without assimilation, while the third and fourth experimental sets run the Latent Heat Nudging data assimilation on a bulk and double-moment scheme, respectively.