Key points are not available for this paper at this time.
An ensemble wave prediction system (EPS) at a 1/16o grid horizontal resolution is being developed for the Mediterranean Sea, based on the WAM Cycle 6 wave model implementation. Open boundary conditions are provided by a North Atlantic wave model at 1/6o resolution. Uncertainty is forced in the Med-waves EPS (in both the Mediterranean and the North Atlantic wave models), through a reduced size ensemble of 10m winds provided from the ECMWF atmospheric EPS (50 members + 1 control run) at approximately 18 km spatial resolution with a forecast horizon of 10 days ahead. The nominal ECMWF EPS ensemble size (50 ensemble + 1 control members) is reduced to 23+1 members using clustering techniques (k-means method). The Mediterranean 10-day ensemble wave forecasts are verified for a 6-month period (Jan-Jun 2022) against available satellite and wave buoys observations, as well as reference cases, like the forecast of the NRT Med-waves of the Copernicus Marine Service. A variety of both deterministic and probabilistic metrics are used in the verification process, in order to access as many forecast quality attributes as possible and to best determine the performance of the EPS. In general, the wave EPS shows a good forecast skill, which after the first 3-4 forecast days becomes even better than the skill of the deterministic Med-waves Copernicus Marine forecast system, indicating the importance of the ensemble approach as the forecast ranges increase.
Kampouris et al. (Fri,) studied this question.