Numerical Modelling of Stratification in Lake Constance with the 1-D hydrodynamic model DYRESM

Abstract

Numerical models assist in the understanding of the complex interactions of physical, chemical and biological processes taking place in a lake. A validated model may be employed as a tool in the management of a lake. For example, the model may be capable to predict the consequences of a climate change, a reduced nutrient load or the fate of a contaminant spilled into the lake. While a three-dimensional model is commonly used for detailed but short simulations of single events, the one-dimensional model DYRESM is used for seasonal or long-term simulations of the vertical salinity and temperature distribution. The accurate simulation of salinity and temperature is a crucial prerequisite to couple a water-quality model to DYRESM that models nutrients, plankton and suspended solids. DYRESM is virtually calibration free and has a very short run time.

Lake Constance (in German: Bodensee) is an important water resource as it supplies high quality drinking water for several million people. The goal of this Master's thesis was to test DYRESM for its applicability for long-term simulations at Lake Constance. Almost all available data required to run and validate DYRESM since 1960 were collected and carefully interpreted. The standard version of DYRESM, however, was not capable to model stratification in Lake Constance with sufficient accuracy. It was observed that the simulated thermocline had much stronger gradients than in reality. A DYRESM code extended with an algorithm that parameterizes internal and benthic boundary layer mixing by two non-dimensional numbers was then applied. The fundamental relationships of the algorithm state that the degree of vertical turbulent diffusion is inversely proportional to the Lake number and that the distribution of internal mixing to benthic boundary layer mixing is dependent on the Burger number. It was then shown that the smooth thermocline frequently observed at Lake Constance can be simulated successfully with the implemented mixing algorithm. Hence, the one-dimensional model DYRESM confirmed the enhanced metalimnetic mixing of Lake Constance that has been observed by other researchers. The sound data base and the successful application of DYRESM enables the continuation of this work by running the water-quality model CAEDYM for Lake Constance.