Assessment of global wave models on regular and unstructured grids using the Unresolved Obstacles Source Term

• Published in: Ocean dynamics 2020-11, Vol.70 (11), p.1475-1483
• Main Authors: Mentaschi, Lorenzo, Vousdoukas, Michalis, Montblanc, Tomas Fernandez, Kakoulaki, Georgia, Voukouvalas, Evangelos, Besio, Giovanni, Salamon, Peter
• Format: Article
• Language: English
• Subjects: 2019; Advection; Atmospheric Sciences; Barriers; Cliffs; Earth and Environmental Science; Earth Sciences; Finite element method; Fluid- and Aerodynamics; Geophysics/Geodesy; Islands; Large-scale models; Monitoring/Environmental Analysis; November 10-15; Numerical schemes; Ocean models; Ocean wave effects; Ocean wave models; Ocean waves; Oceanography; Parameter estimation; Parameters; Scale models; Topical Collection on the 16th International Workshop on Wave Hindcasting and Forecasting in Melbourne; Unstructured grids (mathematics); Wave direction; Wave models
• ISSN: 1616-7341; 1616-7228
• DOI: 10.1007/s10236-020-01410-3

The Unresolved Obstacles Source Term (UOST) is a general methodology for parameterizing the dissipative effects of subscale islands, cliffs, and other unresolved features in ocean wave models. Since it separates the dissipation from the energy advection scheme, it can be applied to any numerical scheme or any type of mesh. UOST is now part of the official release of WAVEWATCH III, and the freely available package alphaBetaLab automates the estimation of the parameters needed for the obstructed cells. In this contribution, an assessment of global regular and unstructured (triangular) wave models employing UOST is presented. The results in regular meshes show an improvement in model skill, both in terms of spectrum and of integrated parameters, thanks to the UOST modulation of the dissipation with wave direction, and to considering the cell geometry. The improvement is clear in wide areas characterized by the presence of islands, like the whole central-western Pacific Basin. In unstructured meshes, the use of UOST removes the need of high resolution in proximity of all small features, leading to (a) a simplification in the development process of large scale and global meshes, and (b) a significant decrease of the computational demand of accurate large-scale models.