"Comparison of different automated state of the art hydrograph separation approaches applied to sub-catchments of the Upper Danube Basin"
PDF-VersionThe work presented is part of the “GLOWA-Danube” project (GLOWA: Global Change of the Water Cycle, www.glowa.org) which is developing the Decision Support System “DANUBIA”. This system is comprised of 15 models to simulate the impacts of Global Change on the hydrological cycle within the 77.000 km² catchment area of the Upper Danube Basin. The individual elements of DANUBIA; a 3D groundwater model, a SVAT-model, and a surface water model, exchange water fluxes via customized model interfaces. The calculated exchange fluxes are not satisfying, especially in the early stage of model development. This is due to each model needing to be run and calibrated individually at first. Later on, when running in coupled mode, boundary conditions and input parameters that were previously calculated or determined individually (often in a simplified way), are read from the output of the partner models. Therefore, a second coupled-mode calibration becomes necessary. Individual and coupled calibration may have to be repeated several times – which is a tedious procedure, particularly for the groundwater model with a high calculation time. It is, therefore, desirable to determine the exchange fluxes independently from the model calculations using reliable, measured data. Long-term daily measured river discharges at existing gauging stations in the catchment provide information on flow components that originate from direct runoff (surface runoff and interflow) and groundwater runoff (baseflow). In order to use this information to gain a better understanding of exchange fluxes in a coupled model system, the runoff components must be separated. In this study, several empirical hydrograph separation techniques were compared, three of which were studied in greater detail and applied to nine sub-catchments (90 to 600 km²) using daily time series from 1980 to 1999. The results are compared spatially and temporally and reasons for regional distinctions are given. The Base-Flow-Index (BFI) was determined for every sub-catchment in order to compare the different hydrograph separation approaches. The BFI is the average annual part of the baseflow on the total runoff and varied between 0.6 and 0.9. Different separation methods showed very similar BFI in every catchment. Differences between the different geological groups were not very clear, showing that other factors also play an important role (slope, land use, soil type etc.). The DIFGA approach (Schwarze, 1991) also provides information on the average monthly delayed and fast groundwater recharge, and therefore allows the determination of the monthly net storage flow – a figure important in coupled hydrological models like GLOWA-Danube. Advantages and disadvantages of chosen hydrograph separation techniques are presented. Ways for providing area-wide information on groundwater runoff and groundwater recharge are proposed. Hydrograph separation using recursive digital filter techniques is the fastest possibility to analyse a great number of sub-catchments and was therefore applied to 80 gauging stations in the Upper Danube Basin. Nevertheless, it is necessary to carry out a regionalisation, e.g. linear regression, kriging or fuzzy cluster analysis, depending on geological, topographical, pedological, climatic, and land use data.