Community and Infrastructure Resilience Prediction and Management in a Changing Climate

Abstract

Climate change poses the most pressing global challenge in recent history. The risks associated with climate change do not only pertain to the rise in temperature, but the accompanying changes in the meteorological and hydrological properties of the planet. Climate change’s impact on the established communities is vast and visible, affecting the functionality of societies on a multitude of ways, and increasingly causing cascading failures and systemic risks (i.e., failures resulting from the interdependent nature of our society systems). This is aggravated by the expansive development of urban areas into exposed, hazard-prone regions. One of the costliest and most frequent hazards resulting from climate change is flood hazard, with its increasing severity and frequency due to the coupling of the aforementioned reasons. This thesis aims at enhancing the resilience of the exposed communities to climate change-induced hazards, with a focus on flood risk, to develop pertinent realistic, proactive, resilience-informed risk management plans. The thesis applies machine learning and data analytics to understand, quantify, and eventually predict climate change-induced flood risk. Descriptive analytics techniques were employed to understand the extent of flood risk on urban communities, resulting in a categorization of the different community responses to flood risk. This categorization is subsequently employed in developing predictive analysis, where global climate models are utilized to predict the changes of the resilience of the exposed communities until the year 2050. Said studies, while revolutionary in nature, serve as a steppingstone in developing a comprehensive, proactive, global disaster management plan. Finally, the thesis narrows its scope by focusing on operationalizing the developed climate resilience methodology considering a single critical infrastructure network and enhances the climate resilience of its risk management plan, set, and operated by its asset owners and decision makers. The approaches developed herein were applied on different datasets for vulnerability identification, loss and resilience prediction, and policy improvement resulting in an overall climate resilience-informed enhancement of the current risk management practices.