ABSTRACT
The aim of this review is to provide an overview of the status of validation of eleven biogeochemical and ecological models of the greater North Sea (COHERENS, CSM-NZB, DCMNZB, DYMONNS, ECOHAM, ELISE, ERSEM, FYFY, GHER, NORWECOM, POLCOMSERSEM) showing the realism achieved as well as the problems hindering a better degree of validity of the models. Several of the models were able to reproduce observations of the state variables correctly within an order of magnitude, but all models are not capable of reproducing every simulated state variable in the range of observations. None of the models can be called a valid model. Comparison of results from different models with datasets are evaluated according to the different spatial and temporal scales, for which data products were available, namely for regional distributions, annual cycles, long-term developments and events. The higher the trophic level, the greater was the discrepancy with the data. Problems still exist in determining the necessary complexity of the ecosystem model. More complexity in the model does not necessarily improve the simulations. Special attention should be devoted to the regeneration mechanisms in the sediments. Species’ groups have been simulated so far with rather limited success. The ecological model simulations did not reproduce fully the observed variability. Possible sources of lacking coincidence with observations originating from the spatial and temporal resolution of the internal dynamics, the trophic resolution, or the resolution of the forcing functions are discussed. Most of the models still need to be evaluated more intensively for their predictive potential to be judged. They have not yet been tested to a degree which is possible today using the various existing datasets from the northwest European shelf seas (presented in the Appendix). Common datasets for the necessary annual cycles of forcing functions are needed.
