Aarhus Universitets segl


Nielsen, A., Andersen, T.K., Søndergaard, M. & Trolle, D. 2020. Dynamiske, procesbaserede modeller som forvaltningsværktøj for danske søer. Forsknings- og udviklingsprojekt vedr. muligheden for anvendelse af dynamiske modeller i forvaltningen af søer i Vandområdeplaner. Aarhus Universitet, DCE – Nationalt Center for Miljø og Energi,56 s. - Videnskabelig rapport nr. 402. http://dce2.au.dk/pub/SR402.pdf


Within the framework of this project, tools have been developed that standardise and partially automate the workflows in relation to setting up lake-specific dynamic models. Thus, models based on the one-dimensional GOTM-FABM-PCLake model have been set up and adapted to five different Danish lakes (Ravnsø, Bryrup Langsø, Søholm Sø, Arreskov Sø and Hinge Sø). The ability of the models to simulate the ecological state of these lakes is compared with results from other dynamic model studies and also with the empirical model approach used in the Danish water management plans.

When the dynamic, process-based models are compared with both Danish and international examples from research articles, the results for the five Danish lakes are very good. The calibrated models represent physical, chemical and biological elements (the latter in the form of chlorophyll a, which is used as proxy to assess the ecological state) in the lakes at a level that, as a minimum, reflects the results of other studies, but in most cases a better correlation with observations is achieved.

The results of the project also show that the uncertainty (in the form of the absolute percentage bias) regarding the description of the state of, respectively, total phosphorus and chlorophyll ranges, on average, from 40% and 54% based on the empirical models presently used in the water management plans to 15% and 22% based on the dynamic models. This is not surprising since the dynamic, process-based models are adapted to specific lakes, whereas the empirical models are built on data from a wide range of different lakes. It should be noted, however, that some of the empirical models may in some cases reflect the average of observations well and even better than the dynamic model, which, for example, was the case with the empirical Vollenweider model for Ravnsø.

Dynamic, process-based models can also be used to simulate higher trophic levels (algae groups, zooplankton, submerged vegetation and fish), which, with the exception of zooplankton, have direct relevance for the biological quality elements used in the water management plans. However, the model output is not always directly comparable with the way in which the biological quality elements have been observed and quantified in the monitoring programme. Therefore, it is recommended that empirical models are developed that combine the biological indices used in water management plans and elements from the dynamic model (e.g. output for chlorophyll a and nutrient levels). This is also the current practice in the management of marine areas, where both dynamic and empirical models are used in an integrated manner.

For a number of Danish lakes, the available data suffice for the use of dynamic, process-based models. When taking into account that these models, in contrast to the simple empirical relationships, can also provide insight into interactions in the ecosystem, the temporal development in ecological state after an intervention in the catchment or lake (restoration) as well as the effect of climate change, this project has shown that there is a potential for using these models to a greater extent in Danish water management.