12.2 Development of Three-Dimensional Unstructured Grid Finite-Volume Model for Coastal and Estuarine Circulation, the General Ocean Model (GOM)

Wednesday, 15 January 2025: 4:45 PM
218 (New Orleans Ernest N. Morial Convention Center)
Jungwoo Lee, East Carolina University, Greenville; and J. Lee, S. L. Yun, and S. K. Kim

We developed a new three-dimensional unstructured grid coastal and estuarine circulation model, named the General Ocean Model (GOM). GOM was developed by combining the finite difference method (FDM) and the finite volume method (FVM), taking advantage of the computational efficiency of the FDM and the exact conservation of the FVM. Significant computational bottlenecks that arise when solving the propagation and the non-linear advection terms are relaxed by using a semi-implicit numerical scheme and the time-explicit Eulerian-Lagrangian method. In addition, the shared memory open multi-processing (OpenMP) was applied to achieve high simulation efficiency. By utilizing orthogonal unstructured computational grids, allowing both triangular and quadrilateral grids, GOM is considerably flexible to resolve complex coastal boundaries without any transformation of governing equations. The GOM was successfully verified with five well-known analytical solutions: (1) Wind Setup, (2) Tidal Propagation with Constant Depth, (3) Tidal Propagation with Nonlinear Advection, (4) Quarter-Annular Harbor Test, (5) Wetting and Drying Test on Tidal Flat. The intrinsic wetting and drying calculation process allows the GOM to be used for coastal flooding and shallow tidal flat studies. Finally, GOM was applied to the Texas coast, including Galveston Bay and Sabine Lake, to show its robustness and efficiency when applied to a relatively large coastal study area. The water surface elevation comparison results at multiple NOAA stations showed an overall skill value of 0.951, and the results were also compared with a well-known SCHISM model results to cross-verify the model’s accuracy and efficiency. The verification results showed that the algorithm used in GOM was correctly coded, and it is efficient and robust.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner