Predictability of wind-induced sea surface transport in coastal areas

In this work we investigated the predictability of the wind-induced sea surface transport in coastal areas. The wind fields predicted by two state-of-the-art meteorological models, namely ECMWF and SKIRON, were used as forcing for a hydrodynamic and particle-tracking model applied to reproduce a set of observed drifters trajectories in a coastal area of the Mediterranean Sea. A set of anemometric data derived by in situ measurements were also adopted as model forcing to reproduce the observed drifter paths. This approach provided a baseline that was used as a reference for evaluating the effects of the predicted wind accuracy on the numerical model solution. The accuracy of the simulation results obtained using, as model forcing, the observed wind data was fair and suitable for most of the operational oceanographic purposes. It decreased when using the wind data predicted by the two meteorological models. In particular, the results obtained using ECMWF data were about 3 times more accurate than the ones obtained using SKIRON ones. The uncertainties were strongly dependent on the range of observed wind speed classes with a different behavior depending on the type of adopted wind data. Finally, the amplification of the errors in predicting the sea surface transport generated by the inaccuracies of the predicted wind fields was quantified.