Abstract:The response of lake systems to changed climatic conditions has important practical implications and is a question of ongoing research. When providing input data for climate impact modelling on lakes, two general uncertainties set the background on which to work on:
(1) Lake system reactions
are non-linear and not fully understood.
(2) Climate models (even regional ones) do not offer the discretization needed and come with great inherent uncertainties. To handle both uncertainties in a reasonable manner, we want to develop a method that (1) allows lake modellers to gain deeper process understanding by defining changed weather themselves and (2) provides stochastically downscaled climate projections. We will develop a novel approach - combining occurrence probability and intensity - to deal with the difficulty of rain simulation and extend an existing single-site weather generator to multi-site simulation. Copulas are planned to be used where conventional measures of dependency are not suitable. Hydrodynamic and ecological modelling will be done on the example of Lake Constance, a lake that serves as a drinking water resource for 4.5 M people.