|SMOES: a simulation model for the Oosterschelde ecosystem: 2. Calibration and validation|
Scholten, H.; van der Tol, M.W.M. (1994). SMOES: a simulation model for the Oosterschelde ecosystem: 2. Calibration and validation, in: Nienhuis, P.H. et al. (Ed.) The Oosterschelde Estuary (The Netherlands): a case-study of a changing ecosystem. Hydrobiologia, 97: pp. 453-474
In: Nienhuis, P.H.; Smaal, A.C. (Ed.) (1994). The Oosterschelde Estuary (The Netherlands): a case-study of a changing ecosystem. Reprinted from Hydrobiologia, vols 282/283. Hydrobiologia, 97. Kluwer Academic: Dordrecht. 597 pp., more
In: Hydrobiologia. Springer: The Hague. ISSN 0018-8158, more
|Also published as |
- Scholten, H.; van der Tol, M.W.M. (1994). SMOES: a simulation model for the Oosterschelde ecosystem: 2. Calibration and validation. Hydrobiologia 282-283: 453-474, more
Calibration; Optimization; Optimization; Optimization; Output; Output; Output; Uncertainty; Uncertainty; Uncertainty; Uncertainty analysis; Validation; Validation; ANE, Netherlands, Oosterschelde [Marine Regions]; Marine; Brackish water
Ecosystem models; Estuary; Estuarium
|Authors|| || Top |
- Scholten, H.
- van der Tol, M.W.M.
SMOES is an invalid model in the Popperian sense of the word. Yet it might be a useful model, which accurately simulates some of the ecosystem behavior. Our knowledge of the Oosterschelde ecosystem is empirical and consists of a comprehensive set of observed data (experiments and in situ measurements). Model usefulness will be investigated in terms of model adequacy (can a model simulate all system behavior) and reliability (does it simulate observed system behavior only).The large impact of the building of a storm-surge barrier in the Oosterschelde Mouth (compartment West) allows us to distinguish two different systems. The quality of the simulation model, built and calibrated for the pre-barrier system, can so be tested by using it for the (new) post-barrier system. The question if the model does predict the post-barrier future will be answered.We apply a calibration method which generates not merely a single best value for each of the many badly known parameters, but which also enables us to estimate the uncertainty in the model outcome. If we use this method to calibrate SMOES for the pre- and for the post-barrier situation, and subsequently estimate the uncertainty in the model results, we find no large (functional) shifts in the behavior of the main ecosystem properties.