Spatial complexity and reservoir optimization results

The complexity of a model can significant;ly affect its cost of development, ease of use, and the reliability of its output. However, it is difficult to find the levels of complexity appropriate for a particular model application because few quantitative studies of the effects of complexity on model results exist. This paper attempts to classify various types of model complexity. Using three indicators of spatial complexity, network flow models of Northern California's water system are formulated and compared at six levels of spatial aggregation. The results show that less complex spatial formulations result in fewer modeled shortages and lesser economic costs. However, the model continues to respond realistically to changes in hydrology. Additional research is needed to more completely understand the effects of model complexity on results, but these results provide an indication of the magnitude of these effects for a large water resource system.