Efficient estimation of directional wave buoy spectra using a reformulated Maximum Shannon Entropy Method: Analysis and comparisons for coastal wave datasets

Directional spectra are a fundamental property of waves, determining how they interact with structures and the wider marine environment. Directional spectra can be estimated (although not directly measured) by wave buoys, the ubiquitous source of in-situ wave and seastate data. The most accurate standard method for estimating the directional spectrum is the Maximum Shannon Entropy Method (MEM-II). However, in practice, speed and numerical convergence issues force engineers to routinely rely on the simpler, but less representative methods such as the Maximum Burg Entropy Method (MEM-I) or one-parameter reconstructions.By reformulating the underlying equations, we have developed a new algorithm (and associated Matlab tool) to accurately and easily generate MEM-II reconstructions, capable of producing a month's worth of half-hourly spectra in ten minutes. A year's worth of directional data for six UK coastal locations was analysed using this method. Spectra were compared with those from the simpler cos 2s and MEM-I methods, which were seen to produce misleading results for real sea data, including large systematic errors in spectral width. The MEM-II reconstruction also offered insights into directional bimodality for both wind seas and swells.