A hydrological model needs to be calibrated to obtain the most optimal parameter set and the most common method is the use of optimization algorithms to minimize the error between observed and modelled values. While it can be achieved by minimization of the sum of squared errors or absolute errors, several other target variables are suggested in various literatures. Among those, Nash–Sutcliffe Efficiency and Kling Gupta Efficiency are commonly adopted, both of which are similar and different in their own ways. Further, parameters such as RMSE, percentage bias, RSR (RMSE-observations standard deviation ratio) etc. are also adopted in many studies as model evaluation statistics due to their respective utilities in various situations. In this paper, impact of choosing different parameters as the objective function is examined. For testing, a lumped hydrological model is applied on two mountainous catchments in western USA. The calibrated parameters obtained from using the different performance metrics as the optimization target are used in simulation to examine their relative merits in terms of the other metrics.