Benchmarking and classification of multi-source metocean datasets in the northern Tyrrhenian Sea

In the context of developing an accurate and extensive knowledge of metocean conditions, this study presents a benchmarking and classification of multiple datasets in the northern Tyrrhenian Sea. The multitude of resources collected within the framework of the AIMS monitoring project, including both in-situ measurements from oceanographic buoys and numerical model outputs, enabled an extensive comparison aimed at assessing the consistency, reliability and representativeness of each data source. Advanced performance metrics, such as Root Mean Square Deviation, Pearson correlation coefficient, were leveraged and combined, providing a unique and comprehensive criterion for data source assessment and classification. While the ERA5 dataset performs poorly in closer basins such as most areas of the Mediterranean Sea, the local SWAN model is able to provide great results in terms of accordance with observed reference, if run with a sufficiently fine grid. Moderate performances can be achieved with a Mediterranean-referred intermediate-scale model such as the CMEMS one.