Modeling Concepts - Coastal Lagoons: Ecosystem Processes and Modeling for Sustainable Use and Development - page 288

complex models (even in their simplest formulation) and to establish advanced data
collecting and data supply programs ( see Chapter 7) is in all cases more progressive
than the use of a restricted simplified model based on the traditional monitoring
data. Model simplicity is very attractive to administrators but modeling progress is
in teaching environmental engineers to use more complex models instead of the
“shortest” but unsustainable application of simple models to a wide range of different
tasks.

6.4 MODEL IMPLEMENTATION

As described in Chapter 3, the implementation of a numerical model to a specific
lagoon requires the specification of the boundary conditions and model param-
eters specific and suitable for the site and for the conditions envisaged in the
simulations.
The bathymetry and the tidal characteristics at the inlet(s) are boundary condi-
tions that can be considered as time independent or deterministic in most sites. On
the other hand, river discharge, water column structure at the sea boundary (in the
case of 3D simulations), and the atmospheric conditions are time-dependent bound-
ary conditions.
Implementation of a model requires data to specify these boundary conditions,
the initial conditions, and the various parameters for process simulation. Additional
data are required to calibrate and validate the quality of model results.

6.4.1 B ATHYMETRY AND THE C OMPUTATIONAL G RID

Bathymetry describes the lagoon's geometry and is the basis of the whole modeling
procedure. Bathymetry is generally measured by national hydrographic authorities
using a fine resolution in order to provide data for navigation charts and to support
various coastal engineering works. In terms of bathymetry and grid definition, laterally
integrated models constitute a particular case, requiring much less information.

6.4.1.1 Laterally Integrated Models

In the beginning of this chapter it was shown how to build a 1D model and the
parameters characterizing the respective grid cross section and cell length. In that
case, the volume of each cell was calculated as the product of the cross section by
the cell length. Generally, the free surface level varies in time and the model
calculates the value of the cross section as a function of that level. Under these
conditions, the information required is the width of the cross section as a function
of the level. This information has to be supplied at the two tops of each cell or at
its center. Intermediate values can be calculated by interpolation.
In 2D laterally integrated models the procedure is similar to that in 1D models.
In this case the difference is that a cross section is calculated for each layer, as a
function of its thickness. In the case of sigma-type coordinates (see Section 6.4.1.2)
the thickness is a function of the surface level for all the layers, while in Cartesian
coordinates only the surface layer has a variable thickness.