Modelling flood plain vegetation based on long-term simulations of daily river-groundwater dynamics

The quantitative prediction of ecological patterns in flood plain areas is a complex task. Variabilities and interactions of hydromorphological and biotic components stand against model limitations concerning input data and process representation. With respect to uncertainties, the coupling of abiotic and biotic data and models, a prerequisite for sustainable management, is a scientific challenge. In this paper, the coupling of models is investigated using a GIS and a digital elevation model in a flood plain area of the German Elbe River. Initially, the habitat suitability for a specific flood plain vegetation type was calculated based on reconstructed daily river and groundwater levels (period 1964-1995), with good agreement compared to a map of observed vegetation types. Subsequently, the calculation was repeated based on stochastic simulations of daily river flow (30 series of 32 years). The results reveal the sensitivity of the predicted habitat suitability against long-term hydrological variability, pointing out important components of the predictive uncertainty.