In water-quality modelling several different directions can be distinguished according to the strategy employed and the disciplinary background used for analysis. A precise classification would be difficult to make, but the manifest difference between transport-oriented and ecology-oriented water-quality models creates at least two obvious groups. In the first case, the description of biological and chemical processes is oversimplified, while in the second, the same applies to transport phenomena. Both approaches determine the level of involvement of the respectively less emphasised phenomena by a priori assumptions. This paper discusses this apparent gap and how to overcome it. No overall procedure is given; instead, a framework is suggested which is based on establishing the relative importances of the various subprocesses determining water quality and proceeding to the corresponding model structure. This should allow the proper combination of knowledge gained from theory and observations, and, furthermore, the elaboration of essential modelling steps such as parameter estimation and model identification. To illustrate, two examples are presented-one concerns heavy-metal pollution of a river, the second involves a lake for which the wind-induced interaction between water and sediment was analysed. In the first situation, a one-dimensional, coupled hydrodynamic-transport-water-quality model for three cadmium compartments was adopted. For the second problem, it was necessary to determine the unknown boundary condition at the lake bottom. This latter was achieved by simplifying the governing transport equation into an ordinary differential equation and introducing some simplifying hypotheses. Given data from regular, intensive observations, the calibration, identification and validation of this model were carried out using the extended Kalman filtering technique.
ASJC Scopus subject areas
- Ecological Modelling