Study how information in the initial state of nonlinear systems diminishes in forecast process as a result of chaos and the use of imperfect models.

Traditionally, efforts in forecasting hydrological, as well as most other phenomena focused on finding the most likely future state. Uncertainties in the initial and boundary conditions, as well as in the forecast model, have often been ignored. Ensemble forecasting offers a practical solution for assessing forecast uncertainty. In recent years, extensive work has been accomplished.