Here we analyze the wave hindcast reliability for a proper description of wave climate in the Mediterranean Sea. To this aim, 6-hourly 35- years ECMWF (European Center Medium Weather Forecast) wave data at 0.7° resolution grid are compared with those provided by means of a meteocean modelling chain operative at DICCA, University of Genoa (http://www.dicca.unige.it/meteocean/) covering a 34-years temporal span at an hourly frequency on a 0.1° resolution domain. Results reveal not negligible differences in evaluating significant wave heights at peaks; in particular the tendency to underrate values in storm sea conditions performed by ECMWF dataset is here evidenced. This behavior turns directly into not-reliable long-term return level estimates for extreme wave analysis, leading to a weak description of wave climate; conversely, a wave climate robust assessment is of primary importance for maritime design.
In recent years raising requirement of reliable wave data sets covering longer periods of time at high spatial-temporal resolution has shifted attention toward wave modeling and development of wave hindcast data. In particular, the availability of robust wave hindcast data is of primary importance in the specific framework of extreme waves analysis for coastal and offshore maritime design, coastal defenses studies and erosion risk evaluation. Wave modeling data quality strongly depends on models capability in representing wave dynamics in its whole complexity. In addition to a proper description of offshore and coastal physical processes (e.g. by means of proper numeric settings, adoption of specific propagation schemes, source terms parameterization), the impact of spatial and temporal resolution on the proper description of local waves properties plays a significant role in defining reliable long-term estimates.
In Cavaleri (2009) an exhaustive overview about wave models capability of properly reproducing the conditions during and at the peak of severe and extreme storms is provided, ranging from the correct description of wave physical processes to the importance of wind accuracy as a relevant factor at the peak of the storms. More precisely, low spatial resolution leads to a poor description of local meteorological forcing features and intensities, which translates into a lacking representation of wave extremes, as well as low temporal resolutions imply possible missing peaks. Wind forcing, especially in areas characterized by strong spatial gradients, plays a crucial role in modeling properly wave fields, especially for the accurate description of significant wave heights, in particular during severe and extreme sea conditions. The necessity of adopting downscaled atmospheric fields leading to wind forcing at high spatial resolution represents a basic component in robust wave modeling and for an accurate description of more severe sea conditions (Cavaleri and Bertotti, 2006). The present work lays in the framework of extreme waves analysis based on hindcast data trying to investigate wave hindcast reliability in extreme sea conditions for long-term estimates. To this aim, two different wave hindcast dataset (DICCA and ECMWF ERA-Interim data) at different spatial and temporal resolution are here employed and analyzed in order to evaluate their reliability in relation to intrinsic original features. To our knowledge, no other works comparing quantitatively any existing wave hindcast dataset and ERA-Interim one are available in literature.
