Uncertainty in wave hindcasts in the North Atlantic Ocean

Accurate knowledge of ocean surface waves is crucial for ship design. With the significant advancements in model physics and numerical resources, recent numerical wave hindcast data has the potential to provide environmental conditions for estimating the wave load in the ship design process. To help estimate extreme wave loads with quantified uncertainty for strength assessments of ship structures, this study aims to quantify model uncertainty in the state-of-the-art numerical wave hindcast products. Model uncertainty was deduced from inter-model comparisons with four numerical wave hindcast products after model evaluations. For model evaluations, wave buoys distributed over the Northwest Atlantic and Northeast Atlantic were used as wave buoy arrays. In general, all four models accurately estimated the significant wave height Hs. However, as the wave condition becomes extreme with the increase of Hs, the model estimations of Hs deviate from the buoy data and differ from each other. For the extreme wave conditions, two wave hindcast products consistently showed better performance based on the buoy observations for both the Northwest Atlantic and Northeast Atlantic. In addition to Hs, the mean wave period Tm02 from the hindcast products are also examined in this study. Although evaluations of Tm02 are not often seen in previous studies, Tm02 is another crucial wave parameter because the ship response is often calculated as a function of Hs and Tm02. Based on careful comparisons, it is found that all the wave hindcast products underestimate Tm02 under the extreme wave conditions of Hs>10m. The joint probabilities of Hs and Tm02 for extreme wave conditions of Hs>10m and 8s