Data assimilation has allowed hydrologists to account for imperfections in observations and uncertainties in model estimates. Typically, updated members are determined as a compromised merger between observations and model predictions. The merging procedure is conducted in decision space before model parameters are updated to reflect the assimilation. However, given the dynamics between states and model parameters, there is limited guarantee that when updated parameters are applied into measurement models, the resulting estimate will be the same as the updated estimate. To account for these challenges, this study uses evolutionary data assimilation (EDA) to estimate streamflow in gauged and ungauged watersheds. EDA assimilates daily streamflow into a Sacramento soil moisture accounting model to determine updated members for eight watersheds in southern Ontario, Canada. The updated members are combined to estimate streamflow in ungauged watersheds where the results show high estimation accuracy for gauged and ungauged watersheds. An evaluation of the commonalities in model parameter values across and between gauged and ungauged watersheds underscore the critical contributions of consistent model parameter values. The findings show a high degree of commonality in model parameter values such that members of a given gauged/ungauged watershed can be estimated using members from another watershed.