High-Impact Low-Probability events and Systemic Resilience: A Network-Based Methodology and its application to the metropolitan city of Venice

High-Impact Low-Probability (HILP) Events are catastrophic events characterised by a lack of precedence and high levels of uncertainty due to their cascading and compounding dynamics. Over recent decades, these events have become more likely as the interconnectedness of social and ecological systems has grown, heightening the risk of cascading failures across sectors and amplifying the impacts of initial triggers. As traditional risk-based approaches often fail in analyzing complex risk interactions, a more holistic methodology is needed to identify cross-sectoral vulnerabilities, common failure points, and strategies to improve systemic resilience. The AGILE project aims to fill this gap by developing a methodology for understanding, assessing, managing, and communicating HILP events with a systemic, risk-agnostic and systemic resilience perspective. The proposed methodology is subdivided into 3 tiers with increasing levels of detail: the first tier is the scoping study, i.e. the mapping of the system critical functions and their relationships, and developing a guideline structure for table-top exercises; the second tier refers to the identification and parametrization of interdependences and feedback loops between critical functions to gauge single-points-of failure; finally the third tier aims to uses network analysis and technological innovations (e.g. artificial intelligence and machine learning) to create an asset-level representation of systemic performance, evaluating flows of information, resources, and energy through the system in real time.

This paper describes the methodological approach under development for the third tier and its preliminary application to the metropolitan city of Venice. The system is conceptualized as a network, with nodes representing key elements that drive the city's functionality, such as transportation infrastructure, communication systems, ecosystems, households, and economic activities, while links represent the dependencies between these components.   The quantification of the links poses significant challenges, requiring a combination of available data, expert input, and, where possible, machine learning and artificial intelligence techniques. By drawing analogies between graph metrics and risk variables, this network analysis identifies key elements that could exacerbate system failure (for example, authority and closeness of a node can be associated with exposure and vulnerability of the corresponding critical function). Simulations are used to explore how risks might propagate through the network, offering valuable insights into potential consequences and strategies for resilience enhancement. Efforts are ongoing to define and refine Venice’s systemic network, with a focus on understanding and addressing vulnerabilities to enhance urban resilience in the face of future HILP events.