The usefulness of forecasting simulations obtained by models is often hampered by the inability to identify the uncertainties and quantify the reliability of model results. Uncertainty in model predictions primarily stems from a number of errors related to the model formulation, such as:
inadequate concept and description of processes and interactions;
inadequate description of spatial and temporal variability;
inadequate description of the state of the system (geometry, initial and boundary conditions, system stresses);
incorrect parameter identification and improper specification of their error bounds.
As a result, in most cases, the efficiency of a model decreases rather than increases as more and more variables and processes are added. So, one of the greatest challenges for future modeling technology is the development of means by which models can explicitly account for the degree of uncertainty introduced by aggregating the different components.
3.2.2 Optimization models
One of the major advances in water resource engineering over the last three or four decades, is the development and adoption of optimization techniques for planning, design and management of complex water resource systems. The analysis of these systems may involve thousands of decision variables and constraints. To overcome problems of dimensionality various schemes have been devised, providing decision alternatives which are optimal in some defined sense and which can be used by water managers to assist their decision making.