Validation of the Copernicus Marine Med-WAV modelled spectrum with available buoy measurements in the Mediterranean Sea

Abstract

The Med-WAV system of the Mediterranean component (MED MFC) of the Copernicus Marine Environment Service regularly provides high-resolution analysis, forecast, and reanalysis of wave products. The Mediterranean Sea Waves Analysis and Forecast product (MEDSEA_ANALYSISFORECAST_WAV_006_017, Korres et al., 2022) has been operational since 2017. The hourly wave spectrum is computed at each model grid point and is discretised using 32 logarithmically allocated frequency bins and 24 equally distributed directional bins. Hourly wave parameters are obtained through the wave spectrum, with spectral parameters significant wave height and mean wave period (spectral moments (0,2) wave period) being continuously validated against satellite altimeter data and buoy measurements. Thus, careful monitoring has contributed to a more accurate representation of the Mediterranean wave system via system upgrades (Ravdas et al., 2018). Αccess to the wave spectrum itself may provide additional information on the sea state, revealing, for example, if it is composed of mixed sea systems. For the Med-WAV system, wave spectra have been available since June 2021 (not part of the Copernicus Marine Service catalogue) and are already used for wave downscaling applications within the Med Sea. Studies concerning comparisons of the modelled spectral shape and in-situ data for the Mediterranean basin are limited to this date. Such an analysis can lead to further parameter validation and contribute to system improvements. In-situ 1-D spectra are available through Copernicus Marine in-situ TAC (2022) (product INSITU_GLO_WAV_DISCRETE_MY_013_045) from buoys deployed in the west part of the basin. The modelled 1-D spectra (following the integration of the 2-D modelled spectrum over all directions) are compared against quality-controlled data from selected deep water buoys. Besides the spectral shape, further comparisons are performed, focusing on parameters that are of interest to the engineering community, e.g. the spectral moments (-1,0) wave period, the spectral moments (0,1) wave period, and the orbital wave velocity (Stopa et al., 2016). The model skill is assessed through commonly used quality metrics such as bias, root mean square difference, and scatter index.