Improving the Numerics of a Third-Generation Wave Action Model
Abstract
A higher order numerical scheme for geographic wave action propagation is chosen as a potential alternative to the existing scheme in the third-generation wave action model SWAN. The chosen scheme is a cyclic hybrid of upwind and centered differencing. The SWAN model with the new scheme (dubbed SWAN-X ) is tested against analytical solutions and experimental (lab and field) data. SWAN-X requires significantly more computational time than SWAN (primarily in stationary mode), since nonphysical oscillations are manifested with high Courant numbers despite the stability of the scheme. It is felt that the advantages of the scheme are best seen in large-scale propagation problems and in wave propagation over rugged bathymetry and/or islands. For most situations over the continental shelf and in nearshore areas, SWAN-X exhibits greater variability than SWAN an indication of reduced numerical diffusion. Artificial diffusion techniques to combat the garden sprinkler effect are discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 03, 1999
- Accession Number
- ADA458637
Entities
People
- Erick Erick Rogers
- James Kaihatu
- Leo Holthuijsen
- Nico Booij
Organizations
- United States Naval Research Laboratory