Critical Infrastructure Systems (CISs) are key for providing essential services and managing critical resources. The failure of one CIS can result in severe consequences on national security, health & safety, the environment, social well-being, and the economy. However, CISs are inherently complex, operating as systems-of-systems with dynamic, non-linear, and uncertain operation conditions, all geared towards fulfilling complex operational objectives. The complexity of both system architecture and operational objectives contributes to challenges in comprehending system-level behavior under normal and disruptive conditions. CISs are also highly exposed to multi-hazard environments characterized by probabilistic behaviors that can impact one or more system components—leading to diverse system failure modes. Understanding the dynamic interaction between hazards and the system response in such environments adds another layer of complexity to CISs safety. Addressing such complexity is crucial and it necessitates thorough investigations to ensure the continuous and reliable operation of CISs. Accordingly, the main objective of this thesis is to develop dynamic resilience quantification approaches for CISs in multi-hazard environments, considering the probabilistic behavior of both the hazard and the system. Given that dam infrastructure is one of the most significant CISs, this thesis employs an actual dam system as a demonstration application for the developed models. Nonetheless, it should be emphasized that the thesis focuses on the generalizability of the developed model to the CISs rather than the specificities related to dam systems, which are adopted herein merely to show the utility of the developed models to complex CISs. Specifically, this thesis first employs a meta-research approach (Chapter 2), using text analytics, to conduct a quantitative and qualitative review of extensive prior research focused on CISs operational safety, considering dam and reservoir systems as one of the key CISs. Such meta-research aims to unveil latent topics in the field and identify key opportunities for future research, particularly in addressing limitations associated with existing risk-based and resilience-based safety assessment approaches for CISs. To overcome such limitations, this thesis (Chapter 3) subsequently developed a coupled Continuous-Time Markov Chain and Bayesian network, facilitating the dynamic quantification of CISs failure risk (propagation of the system's probability of failure with time), considering the temporal variation of uncertainties in system components during operations. Starting from where the risk-based assessment ends (the immediate response of the system at the hazard realizations), resilience-based assessment focuses more on the dynamic system functionality gain/reduction and, subsequently, the system deterioration and recovery rates following hazard realizations. Accordingly, this thesis (Chapter 4) presents a resilience-centric System Dynamics simulation modeling approach capable of representing CISs components, estimating their dynamic system performance, and subsequent dynamic resilience (propagation of the system resilience with time). Such a modeling approach proposes a combinatorial procedure for generating multi-hazard scenarios, encompassing both natural and anthropogenic hazards, where one primary hazard can trigger one or more subsequent hazards. As a result, the developed models can investigate system operations under both single and multi-hazard environments. Furthermore, the coupling between System Dynamics and Monte Carlo simulations (Chapter 5) enables the model to seamlessly incorporate the probabilistic behaviors of both multi-hazard and system responses. The developed approaches can provide the decision-makers with a more detailed system representation that includes probabilistic dynamic system components with multi-operational objectives under probabilistic multi-hazard environments (Chapter 6). Moreover, the developed models can introduce more realistic evaluations for risk-adaptive and mitigation plans in real-time, contributing to more efficient safety assessment plans for the CISs. / Thesis / Doctor of Philosophy (PhD) / Critical infrastructure systems (CISs) play pivotal roles in delivering and supporting the essential needs of our daily lives. However, ensuring the safety of CISs poses layered challenges due to the complexity of their systems and operations, compounded by their susceptibility to multi-hazard environments, all with probabilistic behaviors. Recognizing the criticality and safety obstacles associated with CISs, this thesis introduces dynamics resilience quantification approaches for CISs safety based on a holistic system dynamics representation. The developed models are designed to enhance understanding of the system's performance under multi-hazard disruption conditions, considering the probabilistic behavior of both hazards and system response. Moreover, these models yield resilience-based metrics, allowing for the evaluation of the effectiveness of various risk mitigation plans, which would subsequently lead to more reliable safety assessment plans for CISs. Considering that dam infrastructure is a key CISs, this thesis focuses on the former as a demonstration application to show the developed models’ utility and their efficiency in devising resilience-guided assessment plans for CISs.