Modeling of natural systems typically involves conceptualization and parameterization to simplify the representations of the underlying process. Objective methods for estimation of the model parameters then require optimization of a cost function, representing a measure of distance between the observations and the corresponding model predictions, typically by calibration in a static batch mode and/or via some dynamic recursive optimization approach. Recently, there has been a focus on the development of parameter estimation methods that appropriately account for different sources of uncertainty. In this context, we introduce an approach to sample the optimal parameter space that uses nonparametric block bootstrapping coupled with global optimization. We demonstrate the applicability of this procedure via a case study, in which we estimate the parameter uncertainty resulting from uncertainty in the forcing data and evaluate its impacts on the resulting streamflow simulations.