An application of extreme value theory to the management of a hydroelectric dam.

Assessing the probability of very low or high water levels is an important issue in the management of hydroelectric dams. In the case of the Akosombo dam, very low and high water levels result in load shedding of electrical power and flooding in communities downstream respectively. In this paper, we use extreme value theory to estimate the probability and return period of very low water levels that can result in load shedding or a complete shutdown of the dam's operations. In addition, we assess the probability and return period of high water levels near the height of the dam and beyond. This provides a framework for a possible extension of the dam to sustain the generation of electrical power and reduce the frequency of spillage that causes flooding in communities downstream. The results show that an extension of the dam can reduce the probability and prolong the return period of a flood. In addition, we found a negligible probability of a complete shutdown of the dam due to inadequate water level.