View article

Various climate change effects pose increasing risks to the nation’s infrastructure. Available methodologies address the risk-management problem primarily through cost-benefit analysis frameworks, which evaluate a comprehensive set of protection strategies against a wide range of simulated possible future scenarios. However, due to the substantial climate model uncertainties present over the future planning horizon, such strategies can often lead to less informed policies that might be optimal in an average sense, over the mean of anticipated future scenarios, but cannot offer adaptive solutions based on the actual climate effects evolving in time. To address these limitations, in this research, climate risk mitigation is instead formulated as a decision-making problem within a closed-loop stochastic control-based framework using Markov decision processes (MDP), taking real-time data into account, for evaluating …