Abstract
As we continue to exploit and alter the coastal environment, the quantification of the potential impacts from planned coastal engineering projects, as well as the minimisation of any detrimental effects through design optimisation, are receiving increasing attention. Geophysical fluid dynamics simulations can provide valuable insight towards the mitigation and prevention of negative outcomes, and as such are routinely used for planning, operational and regulatory reasons. The ability to readily create high-quality computational meshes is critical to such modelling studies as it impacts on the accuracy, efficiency and reproducibility of the numerical results. To that end, most (coastal) ocean modelling packages offer tailored mesh generation utilities. Geographical Information Systems (GIS) offer an ideal framework within which to process data for use in the meshing of coastal regions. GIS have been designed specifically for the processing and analysis of geophysical data and are a popular tool in both the academic and industrial sectors. On the other hand Computer Aided Design (CAD) is the most appropriate tool for designing coastal structures and is usually the user interface to generic three-dimensional mesh generation frameworks. In this paper we combine GIS and CAD with a view towards mesh generation for an impact study of the proposed Swansea Bay Tidal Lagoon project within the Bristol Channel and Severn Estuary. We demonstrate in this work that GIS and CAD can be used in a complementary way to deliver unstructured mesh generation capabilities for coastal engineering applications.
Introduction: The Multi-Scale Challenge
Coastal engineering projects, such as tidal lagoons or barrages, present a significant "multi-scale" challenge for computational fluid dynamics (CFD). Researchers must simulate large-scale geophysical features, such as 100km-long estuaries, while simultaneously resolving small-scale man-made structures like 10m-wide turbines or sluice gates.
Standard tools often fall short of this requirement. GIS is excellent at handling large-scale geophysical data like shorelines and bathymetry but lacks the precision needed for complex engineering designs. Conversely, CAD is the industry standard for infrastructure design but lacks a geographical coordinate reference system, making it difficult to place structures accurately within a natural environment. This paper addresses this gap by integrating both into a unified workflow.
Methodology: Integrating GIS and CAD
The authors developed a complementary workflow using a mix of open-source and industry-standard tools to ensure accuracy, efficiency, and reproducibility.
- CAD (qCAD): Precise boundaries of the tidal lagoon infrastructure—including the sea-wall deck, turbine housings, and sluice gates—were extracted and edited based on technical drawings.
- GIS (QGIS): Geophysical data was processed here. The CAD-defined structures were imported into QGIS using geo-referencing utilities, accurately placing them relative to the UK Ordnance Survey shoreline.
- Mesh Generation (Gmsh): An interface between QGIS and Gmsh was used to generate an unstructured mesh. This allows element sizes to vary drastically based on proximity to features of interest.
Mesh Resolution and Scaling
A core insight of this work is the "complementary" philosophy: CAD provides the "what" (the engineered structure), and GIS provides the "where" (the geographical context). To model turbines and sluice gates in a 2D framework, they are represented as "internal boundaries," allowing for sub-grid scale parameterizations of energy extraction.
The mesh resolution is intelligently tied to proximity. By maintaining a 10m resolution within 150m of the shoreline and scaling up to 10km in the open ocean, the model captures complex physics where they matter most without wasting computational power on simpler regions.
Results and Simulation
The study integrated high-resolution datasets, including GEBCO 2014, Digimap Marine (HydroSpatial One), and UK Ordnance Survey shoreline data. The resulting unstructured mesh for the Bristol Channel contained approximately 1.2 million vertices and 2.3 million triangles.
A 5-day hydrodynamic simulation using the Telemac system successfully captured ebb tide velocities and complex flow patterns, such as recirculation regions in the lee of piers near the River Tawe. Qualitative results demonstrated that the mesh was robust enough to handle the high-velocity gradients characteristic of the Severn Estuary's extreme tidal range.
Limitations and Future Directions
While successful, the current model has limitations. The simulation did not include "wetting-and-drying" algorithms for the Severn Estuary's massive tidal flats; instead, bathymetry was capped at -10m to keep the model "wet." Future work will incorporate these algorithms to improve accuracy in shallow areas. Additionally, the authors aim to move toward 3D descriptions of flow around turbines and include sediment transport models to better quantify the lagoon's long-term environmental impact.
Summary updated