This is a two-part work. In Part I, low-cost and representative reduced-fidelity models of two versions of the HYDROTEL hydrological model are constructed, using three types of surrogate models and their combination. Level of representativeness and CPU time ratios between original and surrogate models are evaluated to construct final reduced-fidelity models. Part II of this study focuses on the use of these models within an existing efficient calibration method: the hybrid DDS-MADS optimization approach. Based on this approach, this paper proposes a range of calibration frameworks exploiting reduced-fidelity models. The calibration frameworks are assessed and compared and results demonstrate that exploiting reduced-fidelity models within the hybrid DDS-MADS optimization approach decreases the overall computational time while maintaining the quality of the final solutions. The proposed framework provides a range of tradeoffs between computational time and objective function value. Depending on calibration objectives and optimization constraints, users can select the appropriate one.