Climate change refers to the observed increase in global mean temperature due to the increase in greenhouse gases in the atmosphere because of human activities [1]. A change in temperature triggers changes in climate with effects on virtually all aspects of society and the environment, including water availability, food production, people's health, and the stability of ecosystems [2,3].

A key aspect of the assessment reports of the Intergovernmental Panel on Climate Change (IPCC) is the communication of potential adverse impacts of - and response options to - emerging risks due to climate change. In the Special Report on Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation [4], the IPCC defined climate risk as the interplay between hazard, exposure and vulnerability. Climate risks can arise from potential impacts of climate change as well as human responses to it. These include relevant adverse consequences on lives, livelihoods, health and well-being, economic, social and cultural assets and investments, as well as infrastructures and services, ecosystems and species. The interaction between climate hazards and human responses is not well understood and consequently not well represented in models.

Climate risks result from dynamic interactions between climate-related hazards and the exposure and vulnerability of the affected human or ecological systems to these hazards [5]. More specifically, physical climate risk is a function of the severity and likelihood of climatic impact-drivers such as rising sea level, ocean acidification, hurricanes, droughts, and floods (hazard); the presence of people, infrastructures and ecosystems (exposure); and their predisposition to be adversely affected by the climatic impact-drivers (vulnerability) [3]. Hence, climate adaptation involves either reducing the hazard, the exposure or the vulnerability via planning measures such as constructing seawalls and surge barriers, restoring and preserving natural ecosystems, implementing water-efficient practices, enhancing green infrastructures, upgrading building codes, or developing insurance products [6].

Because the consequences of climatic impact-drivers are most strongly felt locally, climate impact studies often focus on local impacts and local adaptation measures. Moreover, most studies on risks caused by anthropogenic global warming are based on conventional, disciplinary or sectoral risk assessments, for instance on water or food scarcity, infrastructure failure and health impacts. However, the occurrence of past extreme weather events has shown that the consequences of climate change can extend far beyond the initial point of impact and can significantly disturb societies across various sectors and regions, leading to systemic risks [7,8] (see also Fig 1 for an example case). As it is known that social phenomena can be subject to cascading dynamics—where seemingly minor events can provoke qualitative changes in social systems, such as the end of the Cold War, economic crises, and the Arab Spring [9]—climate-driven cascading effects need to be better understood and modeled.

Systemic risk as a concept became first widely used in the aftermath of the 2008 financial crisis when the decline of US housing market prices threatened to lead to a collapse of the strongly interconnected global financial system [10]. In general, systemic risk is associated with cascading impacts that spread within and across systems and sectors (e.g. ecosystems, health, infrastructure, the energy and food sector) via the movements of people, resources, goods, capital and information within and across boundaries (e.g. regions, countries and continents), leading to potentially existential consequences and system collapse across a range of time horizons [11,12].

Anthropogenic climate change can also drive systemic risks. Prominent examples are the Russian heatwave of 2010, the Chinese heavy snow and freezing events of 2008 and the Thai flood of 2011 [13,14]. The Russian heatwave led to a shortfall in cereal production causing a cereal price rise on the global market that led to issues of food affordability in several countries and was associated with food riots in Pakistan and parts of the Arab World. The consecutive heavy snow and freezing rain events that hit southern China in 2008, affected various sectors like electricity, transportation, agriculture, energy, ecology, and tourism, and propagated across 20 provinces. Another example is the Thai flood of 2011 (see also Fig 1) which led to over 800 fatalities and affected about 1.8 million households as well as to the temporary closure of suppliers of essential components to industries in the automotive and electronics sectors [14]. The resulting effects rippled beyond Thailand’s borders reverberating through the global economy with disproportionate impacts to Japan, where estimates suggest that Japanese car makers operating profit was up to 55% lower in 2011 due to the Thailand floods [14].

The evidence of climate impacts cascading across sectors and regions led the IPCC to promote a more comprehensive definition of climate risk in its Guidance Note on the Concept of Risk [5], which entails that climate risk will not only arise from potential impacts of climate change, but also from human responses to it. Particularly, in the context of climate change responses, the Guidance Note defines climate risks as the result of “not achieving the intended objective(s), or from potential trade-offs with, or negative side-effects on, other societal objectives”. This opens opportunities to new and more complex considerations of climate risk as systemic risks. Conceptualizing climate risk as a form of systemic risk allows a better assessment of how climate impacts cascade and spread in ecological and social systems within and across boundaries (e.g., ecosystems, species, countries, and continents) [15]. This cascade of impacts occurs via four main transmission pathways [16]:

These four pathways don’t occur in isolation as there is a multi-level interdependence in the economic, societal, and ecological systems. Targeting risk prevention entails understanding the systems and their underlying vulnerabilities with the goal of preventing cascading impacts and system collapse. These transition pathways and their interdependencies need to be better understood and included in impact models to better analyze systemic climate risks.

Understanding impacts of systemic risk requires a holistic and transdisciplinary view, because not only is the overall risk greater than the sum of individual risks [19], but it can also trigger new events potentially crossing the planetary boundaries with system-destabilizing consequences [20]. The set of nine planetary boundaries defines critical thresholds for maintaining the stability and resilience of the Earth system as a whole, such that global environmental functions and life-support systems remain similar to those experienced over the past ~10,000 years [21]. Several of the nine planetary boundaries have already been crossed, including climate change, biospheric integrity, biogeochemical flows, land use, and freshwater which have a direct impact on agriculture and food security. Crossing planetary boundaries increases chances of hitting earth system tipping points [22]. A tipping point occurs when change in part of a system becomes self-perpetuating beyond some threshold, leading to substantial, widespread, and often abrupt and irreversible impacts. The crossing of planetary boundaries and tipping points creates systemic risks and will compromise the ability of future generations to meet their needs [23].

A way to anticipate tipping points is the establishment of monitoring and early warning systems that detect signals of future changes before these occur [24]. Early warning systems can measure sensitivity between changes in input and output variables, instability indicators, tipping thresholds for transitions, criticality factors and control mechanisms to influence transformation processes. Tipping indicators for early warning have been developed and validated on paleoclimate data. According to the Global Tipping Points Report [22], early warning signals have been detected that are consistent with the Greenland Ice Sheet, Atlantic Meridional Oscillation (AMOC) and Amazon rainforest approaching tipping points. Detecting tipping points via early warning systems can therefore ensure that human activities do not trigger the crossing of even more Planetary Boundaries..

According to the Brundtland report [25], sustainable development is “development that meets the needs of the present without compromising the ability of future generations to meet their own needs”. Intra- and intergenerational justice is essential in the context of sustainable development and sets the foundation of the 17 Sustainable Development Goals (SDGs) and the associated targets which were established by the United Nations in 2015 in the 2030 Agenda for Sustainable Development [26]. The 17 SDGs are: No poverty (SDG 1), Zero hunger (SDG 2), Good health and well-being (SDG 3), Quality education (SDG 4), Gender equality (SDG 5), Clean water and sanitation (SDG 6), Affordable and clean energy (SDG 7), Decent work and economic growth (SDG 8), Industry, innovation and infrastructure (SDG 9), Reduced inequalities (SDG 10), Sustainable cities and communities (SDG 11), Responsible consumption and production (SDG 12), Climate action (SDG 13), Life below water (SDG 14), Life on land (SDG 15), Peace, justice, and strong institutions (SDG 16), and Partnerships for the goals (SDG 17). The SDGs identify sectors and necessary conditions for a fair standard of life and well-being for all. The interdependence of the SDGs, including their synergies and trade-offs, needs to be better understood for a systemic risk analysis.

Systemic effects of climate change and the achievement of the SDGs are interlinked (see Fig 2). Impacts of climate change can undermine the achievability of the SDGs. For example, impacts such as the destruction of critical infrastructures, reduced crop yields, and increased water stress in certain regions could jeopardize the achievement of several SDGs like SDG 1, SDG 2, SDG 3, SDG 10 and SDG 15. These SDGs are instrumental in reducing climate impacts and increasing system resilience, as they would reduce the exposure and vulnerability of the key system elements by e.g. improving the health of people and ecosystems, reducing poverty and inequality, and providing safer infrastructures and better education.

Assessing sustainable development strategies involves synergies and trade-offs between climate adaptation, mitigation, and human development from local to global scales. This is addressed in IPCC AR6 WGII [27] which states that implementing adaptation and mitigation measures must be done such that “multiple benefits and synergies for human well-being as well as ecosystem and planetary health can be realized”. But for this to be achievable, the interdependencies need to be better understood.

Integrated natural and human Earth system models can support decision making and changing course under crisis conditions, however, their complexity might come at the cost of their interpretability and practical usefulness. An integrative framework of human-environment interaction connects climate change, natural resources, human development and societal stability which are coupled through complex pathways, depending on sensitivity to change, e.g., climate sensitivity and conflict sensitivity.

Modeling climate-driven systemic risk to the SDGs combines multiple modeling approaches such as Earth system models and Shared Socioeconomic Pathways, system dynamics models of ecosystems and resources, individual decision models to describe human behavior (optimizing or rule-based agent models), social interaction models to simulate social dynamics, instability and conflict (such as game theory, social network analysis or conflict models), as well as integrated models using artificial intelligence and machine learning [24,28–31].

A promising modeling framework in this context is agent-based modeling (ABM), i.e. the computational study of autonomous agents that can choose among options and adapt their action priorities according to motivation, capability and behavioral rules of optimization, satisfaction and bounded rationality [32,33]. Multiple agents interact with each other, responding to environmental and social change, thus establishing agency to deal with systemic risks and implement SDGs based on reasoning, learning and anticipation. ABMs can be used to describe complex multi-agent patterns of interaction, including environmental conflict and cooperation, human migration, opinion dynamics and network building, formation of coalitions and social norms. Several of these processes combine in micro-macro transitions, which refer to the interconnections between small-scale levels and large-scale levels—each characterized by distinct units—in both natural systems (e.g. turbulent swirls to planetary waves, or river systems to the global hydrological cycle), and social systems (from human beings and communities to nations, global institutions and networks). Transitions occur through processes of aggregation (e.g. individuals form larger social structures) or disaggregation (e.g. breakup of larger units into smaller units) (see e.g., [34–36]). Theories for micro-macro transitions, for instance, in economics and the coupled human Earth system, can help to better understand and anticipate systemic climate risks.

ABMs can support stakeholders and decision makers in developing scenarios of systemic risks and simulations of SDG implementation, including trade-offs and synergies. Key indicators in the feasibility of each SDG are their marginal benefits, costs and risks which can be used to prioritize, compare and select alternative pathways, influenced by governance rules and regulations to facilitate desirable actions and prevent undesirable ones. This is particularly relevant for preventing negative tipping points and inducing positive ones ([23,24,37]). Social network analysis can visualize the dynamic switching and tipping between alternative pathways, such as the spreading and diffusion of social behavior and technical innovations, conflict and cooperation.

Governance of climate impacts needs to move to systemic risk assessments—treating and modeling climate hazards and the SDGs as interconnected components, rather thanin isolation, as is often the case. Climate hazards are already characterized by their intensity, occurrence probability and their spatial and temporal scales. Recently, compound hazards have also been extensively studied (e.g. [38]), and now the SDGs need to be incorporated into these analyses.

This integration will be instrumental for developing vulnerability scores of each SDG and their interdependence with other SDGs. These interconnections could be modeled by a complex network of multiple interacting systems and agents, using network analysis as a powerful tool to understand and categorize types of interactions and their stability. This would enable us to study in more detail how climate extremes can propagate through economic and social systems, such as cascading effects through networks of sectors, infrastructures and supply chains.

This is particularly relevant for the governance of sustainability transitions and climate-neutral transformations of socio-economic sectors which require multiple actors and transnational networks, including regions, countries, businesses and citizens acting across multiple levels, each having their decision procedures for setting targets and implementing actions for the 17 SDGs. To prevent counterproductive crises, disasters and conflicts to undermine the foundations for sustainable transformation, a growing research community is studying the evolution of cooperation and sustainable networks that strengthen agency to implement the SDGs, reduce trade-offs and develop synergies within the planetary boundaries, and thereby reducing systemic climate risk.

In this essay we argued for three main aspects for future research in climate-driven systemic risk: