When creating a numerical groundwater model, the previously defined model space is discretised using finite elements. In SPRING, the finite element method is used due to its high flexibility with regard to the positionally accurate modelling of irregular structures and the possibility of local mesh refinements.
In SPRING, irregular triangles and quadrilaterals in two dimensions, and pentahedrons and hexahedrons in three dimensions, are used as element shapes. The elements can approximate the geometry of the structures to be considered (point, line and surface structures) in a piecewise linear manner due to their variable shapes. The corners of the elements are the nodes of the model.
When discretising three-dimensional models, the horizontal mesh is projected onto the underlying layers and connected to form prisms in order to generate the elements of the individual layers. Geological interfaces can be used as the height of the nodal layers. For numerical reasons, it may sometimes be necessary to further refine layers so that several elements lie exactly on top of each other over the thickness of the layer.
The discretisation density must always be selected with regard to the problem to be solved and the available data. If neither measures are planned in a model nor hydraulic structures are active, the node spacing can be 25 to 50 metres. Depending on the size of the model or the problem, however, a higher resolution or finer discretization can always be useful.
The following table provides reference points for the node spacing of different areas:
Scenario |
Recommended node spacing |
Fountains |
Under 5 m |
Planning area |
5 to 15 m |
Waterbodies (potentially influenced) |
5 to 10 m |
Waterbodies (not influenced/constant) |
10 to 15 m |
Sewerage or drainage systems |
10 to 25 m |
Shipping canals |
20 to 50 m |
Model boundary area |
30 to 60 m |
Special case: of well discretization