Floods caused by heavy precipitation are the most serious natural disaster phenomena on northern shores of Mediterranean. The potential of numerical weather prediction models for shortterm forecasting of heavy precipitation events can be increased and monitored by using
microwave and infrared satellite observations. The University of Wisconsin-Non-hydrostatic Modeling System (UW-NMS) has been used for the numerical simulation of the September 27,1992 Genoa flood event, that caused casualties and produced 420 mm of rain northeast of Genoa over a 8 hour period, with most of the rain falling in a 2-4 hour period. The meteorological
situation from which the event originated, together with the Special Sensor Microwave/Imager
(SSM/I) overpasses at the mature stage of the storm, made the Genoa flood case ideal for the
comparison of observed to model simulated structures.
The simulation has been carried out using three two-way nested grids, confining highr esolution to the flood event; we started with two nested grids (37.5 km and 9.4 km resolution),
and then we added a third (2.3 km resolution) to convective scale. The full explicit microphysics
scheme of UW-NMS model (with rain, cloud droplets, pristine crystals, snow, aggregates and
graupel activated) has been used for the third grid. From the simulation we have built a
conceptual model of the physical mechanisms of the flash floods occurring in the northern
Mediterranean shore. The environmental and microphysical profiles produced by the UW-NMS have been used in a radiative scheme that calculates the multispectral upwelling brightness temperatures (TB's) that would be measured by the SSM/I, forming what is referred to as a cloud-radiation database.
The results have been compared with the available SSM/I observations at the mature stage of the storm, and the usefulness of the observations to the forecast by calibrating the location and time of the precipitation has been investigated. In addition, by looking closely at the microwave upwelling TB's predicted by the model during the evolution of the storm, it has been possible to study the sensitivity of the simulated TB domain to the environmental, structural, and
microphysical features of the simulated storm at the different stages of its evolution. Finally, using profile retrieval algorithms based on cloud-radiation databases produced by the model, we have analyzed the results of precipitation profile retrievals from the SSM/I overpasses. These were then compared to the microphysical profiles produced by the UW-NMS simulation. We will
show that real time satellite retrieval can work with mesoscale model in providing real time diagnostic of heavy precipitation and potential flash floods