Year: 2025
Type: conference-paper
Abstract: Various climate change effects pose increasing risks to our built environment and infrastructure.Existing approaches for managing climate-related implications typically employ a riskbased, cost-benefit analysis that evaluates a comprehensive set of mitigation strategies against a wide range of simulated possible future scenarios.However, the problem is characterized by substantial aleatoric as well as epistemic uncertainties, and cost-benefit strategies can often lead to policies that might be optimal in an average sense, over the mean of anticipated future scenarios, but cannot offer optimal adaptive solutions based on the actual climate effects that eventually evolve in time.To address this challenge, we instead formulate climate risk management and related adaptive infrastructure planning as a formal dynamic framework for sequential decision-making.In particular, our methodology follows a closed-loop stochastic control-based approach using Markov Decision Processes (MDP) and Partially Observable MDPs (POMDPs), taking real-time data into account for evaluating the evolving conditions, selecting the best possible life-cycle actions in time.The developed framework is mainly illustrated here through two coastal adaptation applications.Both applications consider the storm surge and sea-level rise hazards as well as a set of diverse action types, including grey-as well as nature-based solutions.Environmental impacts, such as carbon emissions and sequestration, via the notion of the social cost of carbon, are also accounted for, offering a holistic approach that addresses both the economic and environmental dimensions of coastal flood protection.
Authors: K. Papakonstantinou, A. Bhattacharya, A. Sharma, G. Warn
Institution: Pennsylvania State University
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4411489282