Hindcast based global wave statistics

Abstract

A database of wave scatter-diagrams, describing the joint probability of wave height and wave period worldwide, has been developed. The goal of this work is to provide a practical and representative wave input for the design of structures subjected to waves. This new database is believed to improve significantly over existing alternatives. The paper presents the background of this new dataset. There are three main building blocks. First, an accurate wave data source is essential. We have used a state of the art numerical hindcast dataset here, with extensive calibration and validation by altimeters (IOWAGA dataset, from IFREMER). Second, a database of ship positions has been used to account for the bad weather avoidance. The effect of such routing on statistics encountered by sailing ships is here clearly demonstrated. Third, a flexible statistical model is introduced to fit the raw data. The statistical model aims to smooth out sampling uncertainties. The joint distribution is described by a conditional model, the marginal Hs distribution being modelled by a hybrid distribution, and the conditional period by a split generalized normal law. Eventually, the new scatter-diagrams dataset is compared to one existing alternative, and its limitations are discussed. The final result of this work is a set of wave height and period scatter-diagrams; these are given in the current paper.