Large Sample Basin Experiments for Hydrological Model Parameterization: Results of the Model Parameter Experiment–MOPEX. IAHS Publ. 307, 2006, 220–229.
Regionalization of dynamic watershed response
behaviour
MAITREYA YADAV1, THORSTEN WAGENER1 & HOSHIN GUPTA2
1 Department of Civil and Environmental
Engineering, Pennsylvania State University, Sackett Building, University Park,
Pennsylvania 16802, USA
thorsten@engr.psu.edu
2 Department of Hydrology and Water Resources, University of Arizona, Harshbarger Building, Tucson, Arizona 85721, USA
Abstract Approaches to ungauged basin modelling typically use observable physical characteristics of watersheds (e.g. soil data) to directly infer hydrological model parameters, or they use regionalization methods based on parsimonious hydrological models. A different approach to streamflow prediction in ungauged basins is presented here where, instead of model parameters, the model independent hydrological response behaviour is estimated in the form of streamflow indices, and then regionalized with respect to physical characteristics of watersheds. Therefore, the approach uses a data driven regionalization method (under uncertainty) rather than the common hydrological model driven regionalization method. Ensemble predictions in ungauged basins can then be constrained by limits on acceptable hydrological model behaviour. This study utilizes data from 30 watersheds in the UK. Initial results show that the predictive uncertainty of the model can be reduced considerably through this new approach.
Key words ensemble predictions; hydrograph indices; prediction in ungauged basins; predictive uncertainty; regionalization; streamflow characteristics; watershed response