Abstract
Modelling the thermal structure and heat content of lakes in climate change is getting more and more important as a fundament for ecological lake models. Various studies demonstrated the strong impact of climate change on environmental systems, including the ecosystems of lakes. Anyhow there are several gaps in understanding the complex interaction of atmosphere and hydrosphere as well as in modelling their impact on aquatic ecosystems. When using a hydrodynamic lake model for ecological studies, the careful calibration and validation is an essential condition. The results of the validation process and the following statistical analysis will be presented in this contribution. Afterwards the hydrodynamic model is suitable to estimate future trends in the development of the lakes thermal characteristics, e.g. vertical thermal stratification, Schmidt stability or thermocline shift, which have a significant influence on the aquatic ecosystem. Therefor it is advisable to use data of existing regional climate models, e.g. REMO on the basis of different IPCC emission scenarios. Finally the hydrodynamic model can be coupled with an aquatic ecological model to support the future water quality management. We are working with the one-dimensional hydrodynamic model DYRESM (Centre for Water Research 2010), which is a process based model using a Lagrangian layer scheme. It was established successfully in different investigation areas around the world, in Middle Europe for example at Lake Constance. The input parameters of the model are provided by the meteorological network of the DWD and the Bavarian environmental agency. The calibration covers a period between 2001 and 2007,the validation between 2007 and 2011. As study object we selected the pre-alpine, 83 metres deep, currently dimictic Lake Ammersee, which is located 30 km south west of Munich. This lake was chosen due to the fact that in our opinion it could be representative for many other lakes in the northern foothills of the Alps considering their similar geogenic, climate geographic and limnologic character. There is no other lake in Upper Bavaria where such a large calibration and validation of a hydrodynamic model was achieved. To quantify the current model error when comparing the modelled data with the measured data during model calibration and validation at Lake Ammersee, we calculated the root mean square errors (RMSE), mean absolute errors (MAE) and coefficients of determination in different depths. This statistical analysis we will also present in this contribution.