Testing for the transferability of a water quality model to areas of similar spatial and temporal scale based on an uncertainty vs. complexity hypothesis

The transferability of a water quality model, WASP5, to different reaches of a large river, which vary in both quantity and quality of the surface water is tested. The transferability is set in a theoretical framework based on a hypothesis proposed by Snowling and Kramer (Snowling, S.D., Kramer, J.R., 2001. Evaluating modelling uncertainty for model selection, Ecol. Model. 138, 17-30) stating that as a model becomes more complex in process description, the error between simulations and measurements decreases and the overall model sensitivity increases. This hypothesis is extended to include model transferability by observing the behaviour of the complexity versus uncertainty curves when a model is transferred from one area to another. An overall increase in model error and a decrease in parameter sensitivity are indicative of lack of model transferability. Data from the Saale River, Germany, was used to test the extended hypothesis. The constituents modelled are suspended sediments, chloride and two heavy metals, copper and zinc. The parameters describing sorption of heavy metals to suspended solids or particulate organic carbon proved to be very sensitive to simulation results, hence different complexity levels of the sorption process were implemented. The model was first calibrated and then validated separately for the middle and lower courses of the river. The complexity versus the model error and parameter sensitivities was plotted for the two river courses. Juxtaposition of the two sets of curves shows that the model is transferable between the two reaches when simulating heavy metal transport. Some difficulty arose in capturing a peak of suspended sediments, which is due to exuberant phytoplankton growth. A model restructuring is suggested for simulating the middle course of the Saale River to include the storage (dead) zones in this part of the river, which are havens for phytoplankton growth and potential inoculants of large amounts of algae in the main stem of the river.