Robust Flood Risk Management Strategies Through Bayesian Estimation and Multi-objective Optimization

Management and control of flood hazards, the most frequent natural disaster worldwide, has become a greater challenge due to the increasingly unpredictable precipitation and runoff due to climate change. As many rural areas in Iran are vulnerable to flash floods occurring mainly in the spring, more accurate plans are needed to help reduce the risk of related damage. To address this concern, a robust methodology using multi-objective optimization is proposed, which incorporates the large uncertainties in the modeling parameters defining the risk of flooding. The proposed framework has been implemented in the upper catchment of the Taleghanrood river in the Taleghan district in Iran, which is vulnerable to flooding. The results provide a detailed performance assessment of alternative infrastructure designs, which will help to increase the efficiency of flood management strategies. The optimization uses multi-criteria optimization evolutionary algorithms (MOEA) and Bayesian estimation concepts. The resulting specific design plans, as levees' height increases over a 50-year time horizon, for controlling floods under given scenarios reflect the uncertainty in the parameters.