Risk-based drought early warning system in reservoir operation

This paper introduces a risk-based decision process integrated into a drought early warning system (DEWS) for reservoir operation. It is to support policy making under uncertainty for drought management. Aspects of posterior risk, chances of option occurrences and the corresponding options to given chances, are provided to help decision makers to make better decisions. A new risk index is also defined to characterize decision makers’ attitudes toward risk. Decision makers can understand the inclination of attitude associated with any specific probability through accuracy assessment, and learn to adjust their attitudes in decision-making process. As a pioneering experiment, the Shihmen reservoir in northern Taiwan was tested. Over the simulation period (1964–2005), the expected overall accuracy approximated to 77%. The results show that the proposed approach is very practical and should find good use for reservoir operations.