In the last century a variety of tools has been developed to help understand and predict river morphodynamic processes. Numerical morphodynamic models have become a commonly used tool in river engineering practice. River systems are of a dynamic and stochastic nature and the underlying processes are not completely understood. An imperfect description of physical processes, along with the inability to accurately quantify the model inputs and parameters, leads to uncertainty in morphodynamic predictions. In addition, a natural river system is subject to uncertainties that are inherent to spatial and temporal processes in nature.
For this reason, identifying the uncertainty sources and assessing their contribution to the overall uncertainty in morphodynamic predictions is necessary in order to come to grips with system behaviour. This calls for a stochastic method that enables us to indicate ranges of possible morphodynamic states, their probability of occurrence and the estimation of undesired morphological effects. Stochastic modelling of river morphology and its potential in present-day river management practice is the topic of this thesis.