Regional frequency analysis of extreme waves in a coastal area

This study analysed the wave data from several locations in a coastal region to identify the areas with comparable wave height statistics and to estimate regional extreme significant wave heights of designated return periods. The regional frequency analysis algorithm executed in this work utilised the significant wave height data from 35 sites (35°-45°N, -6.5°--11.0°W) in a coastal grid (0.25° × 0.25°) off Portugal in the North Atlantic Ocean, extracted from 44 years HIPOCAS hindcast wave database. Regional frequency analysis algorithms based on L-moments identified the discordant sites (discordant site statistics are substantially diffused from the group of sites statistics), assisted in the formation of homogeneous regions by cluster analyses (data vectors include site characteristics and site statistics) and consequently selected appropriate regional frequency distributions by Z goodness-of-fit test statistic at 90% level of significance to estimate regional extreme significant wave height quantiles of designated return periods. The regional frequency analysis algorithm recognised in certain cases, neighbouring sites as members of diverse regions, revealing that geographical proximity of sites are not ascertaining factors to form homogeneous regions. At-site analyses along with regional frequency analysis enabled to apprehend the precision of the regional extreme quantile as genuine feature of its at-site extreme quantiles.