Systemic risk: Opportunities for inter- and transdisciplinary science to improve societal resilience

Abstract

Systemic risks will be amplified due to global warming as climate-related shocks, like extreme weather events, cause cascading and compounding impacts across sectors and systems. To better understand and manage systemic risk, inter- and transdisciplinary collaborations between scientists and decision-makers are urgently needed. This backstory summarizes contributions to a focus collection showcasing scientific advances in modeling systemic risk and its drivers and provides examples of using local knowledge, sectoral data, and artificial intelligence for a better understanding of systemic risks.

Above image: Illustration of the complex interconnection between climatic impact-drivers (like heat, frost, drought, flood, storms, etc.), actors, sectors, and sustainable development that are linked to and affected by climate-driven systemic risk. Adapted from Ciullo et al.²

Many widely used models fail to capture the dynamic interactions across socio- ecological and technological systems that underlie systemic risk propagation […] we need a new scientific agenda that is interdisciplinary, transdisciplinary, and co-produced with societal actors.

More inter- and transdisciplinary research is needed into systemic risk with close involvement of relevant actors to deepen our understanding, improve our modeling capability and the governance of systemic risk.

Main text

Beginnings

Understanding and addressing systemic risks in an interconnected world

In an increasingly interconnected world, societies are exposed to risks that transcend sectors, national borders, and conventional categories of hazards.^1,2 From climate-driven disasters and biodiversity loss to geopolitical instability and global health crises, today’s risks are not isolated events but part of a broader web of interacting vulnerabilities. The complex nature of risk was also recognized in the sixth assessment report of the Intergovernmental Panel on Climate Change (IPCC 2023).³ This complex, uncertain, and cascading nature of risk that is capable of disrupting entire systems and challenging the foundations of resilience is also referred to as systemic risk⁴ (see Image 1).

Conventional risk governance tools and scientific assessments struggle to keep pace with this new reality. Most are designed to address well-defined, localized, and linear risks. Systemic risks, by contrast, are characterized by non-linearity, feedback loops, tipping points, and long causal chains that are difficult to anticipate and even harder to manage. This discrepancy points to a fundamental mismatch between the nature of the risks we face today and the tools we use to understand and address them.

Scientific approaches to systemic risk are still in their infancy. Current methods are often constrained by disciplinary silos, limited integration across domains, and insufficient treatment of uncertainty. Moreover, many widely used models fail to capture the dynamic interactions across socio-ecological and technological systems that underlie systemic risk propagation. Issues of data interoperability, use or availability of suitable indicators, and limited interactions with practitioners further constrain the usability of scientific insights for decision-making.

To address these gaps, we need a new scientific agenda that is interdisciplinary, transdisciplinary, and co-produced with societal actors. This includes developing and applying tools that can capture complexity and uncertainty (such as systems mapping, scenario analysis, and agent-based modeling), enhancing the usability and interoperability of data and models (e.g., by use of artificial intelligence), and embedding user engagement to ensure legitimacy and uptake of scientific findings.

This focus collection is part of the activities of the Knowledge Action Network on emergent risks and extreme events (Risk KAN) and brings together six contributions that critically assess the current state of systemic risk science and propose advances to better support governance and decision-making. The collection explores methodological innovations, conceptual frameworks, and applied approaches that together lay the foundation for a more integrated and actionable science of systemic risk.

Research methods and interdisciplinary science

Insights on systemic risk methods and modeling

Methodological and modeling approaches to systemic risk analysis

Two papers look at application and method advancement for addressing systemic risk in specific sectors. In Rindsfüser et al.,⁵ the authors conducted a systematic review of 111 studies to analyze how flood risk evolves over time, considering factors, such as hazard, exposure, and vulnerability. They found that diverse methodologies and drivers influence flood risk and highlighted that risk evolution is highly context-dependent and often driven primarily by exposure and land-use changes, with climate change playing a significant but variable role. This paper highlights the need for systematic, model-based flood risk monitoring that periodically assesses and integrates evolving risk components to support adaptive flood risk management and proactive decision-making.

Building on the importance of modeling approaches, Mühlhofer et al.⁶ investigate how extreme weather events trigger cascading impacts and disruptions in critical infrastructure systems worldwide. Using open-source data, they developed a climate impact modeling chain that integrates hazard, exposure, and vulnerability data with a network model. This approach quantified service disruptions across critical infrastructures, including roads, schools, hospitals, power grids, and cell towers, caused by over 700 hazards in 30 countries. These methodological innovations demonstrate how data integration can enhance understanding of infrastructure vulnerability.

Bertolotti et al.⁷ contribute to systemic risk research by introducing a sustainability investment game that serves as both an agent-based model (ABM) and a discrete dynamical system. The model explores the long-term consequences of investment choices in three interconnected sectors: renewables, non-renewables, and the military. The innovation lies in its ability to map agent decision-making onto systemic dynamics, thereby creating a hybrid modeling framework that combines strategic interaction among agents with global system feedbacks. This approach draws on traditions of system dynamics while also incorporating game-theoretical and agent-based reasoning.

Complementing these modeling efforts, Mehryar et al.⁸ conducted a literature review and expert interviews to evaluate the use of artificial intelligence (AI) in climate change adaptation. They examined the application of various AI methodologies across key areas: risk and resilience assessment, policy appraisal, and measure implementation. Their findings revealed that AI is most commonly applied to risk and resilience assessment, with machine learning methods—such as random forests, artificial neural networks, and deep reinforcement learning—being the most frequently used.

Building on these methodological foundations, understanding the broader context of systemic risk, particularly in relation to climate change, is essential for developing effective mitigation and adaptation strategies.

Understanding and managing systemic risks in a changing climate

Hochrainer-Stigler et al.⁹ discuss the interaction between systemic risks, climate change and other risk drivers. They argue that the increasing global interconnectedness due to, for instance, global supply chains and global travel, lets local impacts due to, for instance, extreme weather events or infectious disease outbreaks, cascade globally. To better anticipate, mitigate and manage such systemic risks, they argue that global risks analyses must be coupled with local knowledge and preventative actions. While systemic risks can be global, they usually are triggered by local events. They propose a process for systemic risk analysis by identifying system boundaries and system factors and applying a risk-layering to differentiate between individual-level risks and connectedness risks.

Santos et al.¹⁰ present an integrative framework for systemic risk analysis that explicitly addresses the challenges of interdisciplinary integration, sectoral fragmentation, and multi-scale dynamics. By employing a translator agent-based model, the authors bridge top-down and bottom-up data sources and facilitate the exchange of information across scientific domains. Applied to the case of COVID-19 in Brazil, the approach reveals how exposure and vulnerability differ significantly across social classes. The framework demonstrates how feedbacks between environmental and social systems can be traced more effectively and highlights the need for overcoming disciplinary siloing to account for cross-sectoral impacts and complex risk propagation.

These perspectives highlight the importance of integrating local insights into global analyses as well as the importance of adopting interdisciplinary frameworks.

Challenges and opportunities

Lessons learned and future direction for systemic risk research

The stepwise methodology proposed by Hochrainer-Stigler et al.⁹ combined with knowledge co-production can provide entry points to address risk interactions for transformative changes to reduce systemic risks. Rindsfüser et al.⁵ emphasize that flood monitoring needs the systematic evaluation of flood risk by regularly reevaluating the flood risk factors and their evolution, which could benefit from a more systemic approach as proposed by Hochrainer-Stigler et al.⁹ to also acquire insights and create opportunities to link to other sectors. Mühlhofer et al.⁶ revealed that even areas not directly affected by hazards could experience significant disruptions, particularly in interconnected systems, such as healthcare and education, which depend on power, transportation, and communication networks. Furthermore, failure cascades were found to disproportionately affect poorer areas, with service disruptions exceeding direct impacts by up to ten times. These findings highlight the urgent need for systemic climate adaptation, prioritizing the resilience of interconnected infrastructure networks to mitigate cascading failures, and service disruptions.

As Meyryar et al.⁸ found that AI applications have predominantly focused on hazard and exposure analysis, with less emphasis on resilience and vulnerability due to limited socioeconomic data: There could be an opportunity for further research on the use of AI to curate and integrate vulnerability data in systemic risk assessments. While some studies use AI for policy analysis, they primarily assess past impacts rather than simulating future outcomes, co-benefits, or maladaptation risks, highlighting the need for more forward-looking research. Additionally, a general lack of expertise, capacity, and knowledge in AI and climate or sustainability science within organizations presents a significant barrier to integrating these fields for systemic risk analysis and decision-making. By integrating agent-based decision-making into a structured dynamical system, Bertolotti and Roman⁷ offer a novel tool to explore systemic risks emerging from collective behavior and sectoral interdependence by combining these with agent-based modeling that could be further enriched by using AI-based approaches as suggested by Meyryar et al.⁸ Finally, the integrative framework for systemic risk analysis proposed by Santos at al.¹⁰ demonstrates how feedbacks between environmental and social systems can be traced more effectively and highlights the need for overcoming disciplinary siloing to account for cross-sectoral impacts and complex risk propagation. Implementing such a framework would allow for promoting evidence-based and context-sensitive policies essential for fairer and more effective adaptation and resilience to systemic risks.

The contributions to this focus collection show that more inter- and transdisciplinary research is needed to better manage, adapt to and mitigate systemic risk. Close involvement of relevant actors in this research is crucial to deepen our understanding, to generate actionable knowledge and to improve the governance of systemic risk. As also highlighted in Ciullo et al.,² the main transmission channels of climate-driven systemic risks are closely linked to the sustainable development goals (SDGs) and can challenge their achievement. In turn, sustainable development guided by the SDGs, can contribute to reducing key vulnerabilities and increase societal resilience toward systemic risks.

Acknowledgments

The authors thank Eilif Ursin Reed (CICERO) for support with the figure. JSi and JSch acknowledge funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy—EXC 2037: “CLICCS—Climate, Climatic Change, and Society”— Project number: 390683824, contribution to the Center for Earth System Research and Sustainability (CEN) at the University of Hamburg, Germany. CF was supported by the Institute for Basic Science (IBS), Republic of Korea, under IBS-R028-D1.