Differences in the parameters and calibration efficiencies of two HBV model types

Accurate modeling of discharges in catchments plays important role in solving a large variety of water management tasks. Rainfall-runoff models are widely used for the estimation of the hydrological phenomena such as runoff, soil moisture, and snow water equivalent, etc. Three basic error factors may affect modeled outputs: quality of input data, parameter uncertainties, and model structure. This study is focused on a comparison of the performance of the lumped and semi-distributed version of the conceptual rainfall-runoff TUW model, which represents two different model structures. We focused on how the model structure can affect the parameters and the runoff model efficiencies. For that purpose, we select the 180 Austrian catchments with different morphological and geographical characteristics. We analysed the variability of efficiencies and parameters of both the HBV models types which are calibrated on discharge in the period from 1991 to 2000. As a result, we can conclude that the semi-distributed version of the HBV type model performs better, with the lower spread of the parameters and better runoff model efficiencies in both Lowland and Alpine catchments types.