Sea level forecasting in Venice through high resolution meteorological fields

This paper is concerned with the sea level forecast in Venice (Italy) and aims to improve the prediction of `high water' events that periodically cause the flooding of the city. A numerical finite element hydrodynamic model was used to simulate the sea level in the Adriatic Sea during the months of November 2001 and 2002, in which flooding events occur. The model was run over two different spatial domains: the Mediterranean and the Adriatic Sea. Two kinds of meteorological fields were imposed as forcing to the model: the ECMWF pressure and wind fields, currently used in the operational sea level forecast but often underestimated over the Adriatic Sea, and the high resolution pressure and wind fields computed by the limited area model LAMI. The performance of the hydrodynamic model driven by the different forcing has been quantified in terms of correlation coefficient, root mean square error and bias. The results indicate that LAMI winds, more intense than ECMWF winds and better describing the mesoscale features of the wind field, produce an improvement in the sea level prediction, in particular during high water events. Moreover imposing observed sea levels as boundary conditions of the Adriatic Sea model increases sensibly the correlation with observed data.