Explicit simulations of stream networks to guide hydrological modelling in ungauged basins
- 1Institute of Hydrology, University of Freiburg, 79098 Freiburg, Germany
- 2Institute of Environmental Engineering, ETH Zurich, 8093 Zurich, Switzerland
Abstract. Rainfall-runoff modelling in ungauged basins is still one of the greatest challenges in hydrological research. The lack of discharge data necessitates the establishment of new innovative approaches to guide hydrological modelling in ungauged basins. Besides the transfer of calibrated parameters from similar gauged catchments, the application of distributed data as a hydrological response in addition to discharge seems to be promising. A new approach to guide hydrological modelling based on explicit simulation of the spatial stream network was tested in four different catchments in Germany. In a first step we used a simplified version of the process-based model Hill-Vi together with regional climate normals to simulate stream networks. The calculation of gravity driven lateral subsurface and groundwater flow is used to identify patterns of stream cells, which were compared to reference stream networks and their degree of spatial agreement was evaluated. Significant differences between good and poor simulations could be distinguished and the corresponding parameter sets relate well with the hydrogeological properties of the catchments. The optimized parameters were subsequently used to simulate daily discharge using an observed time series of precipitation and air temperature. The performance was evaluated against observed discharge and water balance. This approach shows some promising results but also some limitations. Although the model's parsimonious model structure could be further improved regarding discharge recession and evapotranspiration, the performance was similar to regionalisation methods. Stream network modelling, which has minimal data requirements, seems to be a reasonable alternative for model development and parameter evaluation in ungauged basins.