A multi-criteria meshing method applied to a shallow water model

In this study a multi-criteria meshing method for a 2D shallow water model is tested for a domain encompassing several spatial and hydrodynamic scales, from the Western Atlantic Ocean shelf in front of the Iberian Peninsula to the Lima River (Portugal) and its estuary. For the mesh design, a series of scalar meshing criteria are identified from known physical factors of the relevant processes being studied, such as the bathymetry, the topographical length scales and the expected characteristics of the flow motion. Computational constraints are applied as well. Each criterion is expressed in terms of a node spacing function. The final node spacing function driving the meshing process is obtained by merging the requirements of the previously defined criteria. An advancing front mesh generator performs the discretization in a single, unstructured and well graded mesh, having element side lengths varying over three orders of magnitude (from tens of km to meters). A shallow water model application tests the generated mesh, where tide elevations and phases are imposed at the ocean boundary and freshwater inflow at the upstream boundary. The model validation shows a general good agreement between computed and observed water elevations, currents and the asymmetry of the tidal signal, with an underestimation of the high water levels for the most upstream station. The wet and dry process is reproduced realistically, as the position of the wetted front matches closely the submerged areas of time-referenced aerial images.