Regionalization for uncertainty reduction in flows in ungauged basins

The objective of this study is to contribute to the reduction of predictive uncertainty in flows in ungauged basins through application of a regionalization method to 56 well-gauged basins in the United Kingdom. The classical approach of regionalization is adopted, where regression relationships between calibrated hydrological model (HBV) parameters and physical characteristics of the basin are established and used to estimate parameters for ungauged basins. The calibration resulted in optimum parameter sets for 48 basins of which 17 were used for the regression analysis. The results of the regression analysis showed statistically significant relationships for six out of seven parameters, and hydrologically sensible relationships for only three parameters. The validation for eight basins revealed that the regionalization model does not perform satisfactorily. The use of default parameter values seems to favour the use of regionalized parameter values. Therefore, the applicability of the classical approach of regionalization for HBV aiming at simulating all aspects of the hydrograph is questione