Comparison of techniques for the optimization of conceptual hydrological models

The author compares the performances of three optimization algorithms as concerns their efficiency in parameter estimation and minimizing the criterion function in the field of rainfall-runoff modeling. This modeling has been put within the framework of nonlinear regression analysis; some results relating to such a formulation are given. Particular emphasis is given to the construction of a daily rainfall-runoff model with continuous derivatives of the criterion function with respect to model parameters.