Overview

Risk management process to deal with natural hazards (e.g. storm, landslide, earthquake, etc.) involves decision making to find an optimal loss-mitigation approach. A good, representative decision analysis process relies on reasonably capturing all possible scenarios that would happen in a natural hazard event and assessing the probabilities. This requirement for the evaluation of possible scenarios exists while there are multiple sources of uncertainties in such a problem and our knowledge of extreme events is limited. In our research, we showed that using the historical data and extending past knowledge to predict a future event is not completely applicable and reliable. Thus, it is important to consider the possibility of irrelevancy of historical data to the problem. The research shows that to consider extreme and rare events, it is required to use a non-informative prior probability in the Bayesian analysis. Such non-informative prior probability distribution is obtained by maximizing the entropy of possible events, which maximize the lack of information. This presentation discusses the motivation of our research to develop a new decision-based framework in order to reasonably capture sources of uncertainties in the natural hazard risk management decision analysis.