Assimilation of satellite-derived rainfall estimates to improve numerical simulations of medicanes

Mediterranean cyclones can occasionally attain characteristics typical of tropical cyclones. Although these cyclones, known as medicanes, occur only few times per year, they have received a growing interest, especially being associated with severe impacts on the densely populated coasts of the Mediterranean. Satellite products allow to monitor the evolution and to identify key features of intense cyclones, and they also provide valuable precipitation estimates over the sea where cyclone develops and observations are usually scarce. The present study aims at investigating the potential of satellite rainfall estimates to improve numerical simulations of medicanes, through assimilation in a high-resolution, convection-permitting meteorological model. The assimilation is based on a nudging technique that progressively modifies the model specific humidity profiles according to the comparison between observed and simulated rainfall. This physically based procedure proves able to correct medicane simulation, inducing a modification in the diabatic processes (latent heat release) that represent a key factor in cyclone lifecycle. Medicane trajectory and intensity are positively modified, especially when an upscaling of the local moisture correction produces an impact on large scale dynamics. These findings are obtained adopting as testbed two recent high-impact events, Ianos and Apollo storms.