Development of 2D Unstructured Meshes Using a Sizing Function Derived
from Euclidean Distances to Coastal Features for the NWM Hydrodynamic
Engine (D-Flow FM) Model
Abstract
Generation of 2D meshes with reduced number of elements while yielding
accurate results is a major challenge in coastal numerical models.
High-quality 2D unstructured meshes were generated using sizing
functions, which were computed from Euclidean distances to coastal
features at given spatial locations and assigned element sizes based on
calculated distances. The coastal features consist of National Water
Model (NWM) streamlines, National Hydrography Dataset (NHD), NOAA Medium
Resolution Shoreline and bathymetric features from the United States
Army Corps of Engineers (USACE). This approach allows improved
integration of the hydrodynamic D-Flow Flexible Mesh (D-Flow FM) model
into the hydrological NWM and results in an optimum number of
computational points. The method grants the user flexibility to control
element sizes and avoids manual iterative procedures by determining an
optimal element-sizing function that defines small element scales in
regions where geometrical and physical characteristics exist, with
larger scales elsewhere. Newly created continental-scale meshes on the
Atlantic Ocean, Gulf of Mexico and Pacific Ocean coastlines demonstrate
the application of the proposed method for automatic generation of
unstructured, high-quality 2D meshes.