Hydrological model parameters space during calibration

Abstract

Hydrological models are widely used for different water resource problems. There are several objective functions which can be used to calibrate a model. In this study, an attempt has been made to analysis the parameter space mapped by different objective functions during calibration. A conceptual hydrological model was calibrated using robust parameter estimation algorithm with different objective functions. It was found that different objective functions have mapped the parameter spaces differently with no common intersection of parameter space. The diagnosis of parameter space has lead us to develop a hierarchical optimisation (HOP) calibration technique. In HOP calibration, the first step is to constrain the search domain using one objective function. In second iteration it further constrains the space using another function. This process is repeated until dynamics of the hydrograph are well represented. HOP-based calibration outperformed any calibration based on single objective function calibration.