Paper presented at: Headwater Control VI: Hydrology, Ecology and Water Resources in Headwaters, IAHS Conference, Bergen, Norway 2005 1 Integrated research on the hydrological process dynamics from the Wilde Gera catchment in Germany Peter Krause and Wolfgang-Albert Flügel 1 Abstract The “Wilde Gera” is an interflow and snow melt dominated highland headwater catchment in the Thuringian Forest in eastern Germany. Inside the basin the scientific test sites “Schmücker Graben” and “Steinbach” are located in which hydrological research is carried out since more than forty years. Due to the long ranging research a comprehensive amount of data for hydrological system analysis, process research is available and was used for hydrological modelling with various models in the last ten years. These modelling results revealed important insights into the runoff generation processes, which have been further investigated and validated with the help of the distributed hydrological modelling system J2000 recently. The model was applied in different spatial resolution to investigate the influence on model performance as well as on data, parameter and conceptual uncertainties and sensitivities. J2000 was calibrated with discharge time series at the basin’s outlet, and multi response validation was car- ried out by means of time series from recording gauges inside the catchment and additional hydro- meteorological data sets, e.g. snow water equivalent and groundwater levels. The findings under- pinned the dominant importance of the fast subsurface runoff component for the catchment’s runoff generation. They, however clearly indicated the importance of research issues concerning the control- ling function of periglacial hillslope sediments and their internal structure on the hydrological process dynamics still unresolved in a regional context. The paper is presenting first results from this research effort and the application of the J2000 hydro- logical modelling system. Special emphasis will be given on model and parameter uncertainties and the definition of interdisciplinary research interfaces for a better understanding of the hydro- geomorphic process dynamics controlling the runoff generation in highland headwater catchments. 2 Introduction As a first approach for an interdisciplinary research project under preparation the J2000 model was set up and calibrated on existing stream flow records as a baseline for the further research. For this pur- pose three different distributions (2500, 614 and 300 HRUs) were used. The existing time series (1.11.89 to 31.10.2000) was split into a calibration (1.11.89 – 31.10.1993) and a validation (1.11.93 – 31.10.2000) period. Model calibration was only performed on the runoff at the catchments outlet. Ad- ditional data was used later to check if the internal process representation works correctly. The best results were obtained with the distribution of 614 HRUs which represents a trade-off between model performance and the representation of the basin’s variability. This paper will only present the results of this middle resolved distribution. 3 The J2000 modelling system The development of the J2000 modelling system started in 1997 as a PhD project for process oriented hydrological modelling in large river basins. The reason for the development of a new system was that the distribution concept of the Hydrological Response Units (Flügel 1995), which was selected for the project work, had not been adapted to large scale basins at these times. The model was successfully applied in three large subbasins (Mulde, Unstrut and Schwarze Elster) of the river Elbe in eastern Germany, including a study about landuse change on runoff generation (Krause, 2001). A thorough review which was carried out after the first applications revealed that the J2000’s systematic concept was not flexible enough to use the system in different environments or for different purposes or scales without further development. In particular, the separation between the system core and the process modules was not perfect in the first version. These limitations led to a new implementation of the whole system within the NetBeans [www.netbeans.org] environment in Java, which makes the J2000 system independent and much more flexible. The precursor to the new development were the possi- bilities provided by the Modular Modelling System MMS (Leavesley et al. 1996) and the Object Mod- eling System OMS (Ahuja et al. 2004) in such a way that the J2000 can be places between these two modelling frameworks. On the one hand it was designed as a modelling framework for hydrological